bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Polypyrimidine Tract Binding are essential for B cell development

Authors:

Elisa Monzón-Casanova1,2*, Louise S. Matheson1, Kristina Tabbada3, Kathi Zarnack4,

Christopher W. J. Smith2 and Martin Turner1*

Affiliations:

1Laboratory of Lymphocyte Signalling and Development, The Babraham Institute,

Cambridge, UK;

2Department of Biochemistry, University of Cambridge, Cambridge, UK;

3Next Generation Sequencing Facility, The Babraham Institute, Cambridge, UK;

4Buchmann Institute for Molecular Life Sciences, Goethe University Frankfurt, Frankfurt

am Main, Germany

*Corresponding authors:

Elisa Monzón-Casanova ([email protected]) and Martin Turner

([email protected])

1 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Abstract

During B cell development, recombination of immunoglobulin loci is tightly coordinated

with the cell cycle to avoid unwanted rearrangements of other genomic locations. Several

factors have been identified that suppress proliferation in late-pre-B cells to allow light

chain recombination. By comparison, our knowledge of factors limiting proliferation

during heavy chain recombination at the pro-B cell stage is very limited. Here we identify

an essential role for the RNA-binding Polypyrimidine Tract Binding Protein 1

(PTBP1) in B cell development. Absence of PTBP1 and the paralog PTBP2 results in a

complete block in development at the pro-B cell stage. PTBP1 promotes the fidelity of the

transcriptome in pro-B cells. In particular, PTBP1 controls a cell cycle mRNA regulon,

suppresses entry into S-phase and promotes progression into mitosis. Our results

highlight the importance of S-phase entry suppression and post-transcriptional

expression control by PTBP1 in pro-B cells for proper B cell development.

2 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Introduction

During development B cells first rearrange VDJ gene segments of their immunoglobulin

heavy chain (Igh) locus and subsequently VJ gene segments from their immunoglobulin

light (Igl) loci to generate antibody diversity (Supplementary Fig. 1a). A burst in

proliferation at the early-pre-B cell stage occurs once developing B cells have successfully

recombined their Igh locus (Herzog et al., 2009). V(D)J recombination occurs during

G0/G1 phase of the cell cycle (Schlissel et al., 1993) and RAG proteins are degraded upon

entry into S-phase to suppress V(D)J recombination (Li et al., 1996). Therefore,

alternation between proliferative and non-proliferative stages during B cell development

is precisely controlled (Herzog et al., 2009; Clark et al., 2014). Several factors have been

identified that suppress proliferation in late-pre-B cells and allow Igl recombination

including interferon regulatory factors-4 and -8 (Irf4, Irf8) (Lu et al., 2003), Ras (Mandal

et al., 2009), Ikaros and Aiolos (Ma et al., 2010), BCL-6 (Nahar et al., 2011), dual specificity

tyrosine-regulated kinase 1A (DYRK1A) (Thompson et al., 2015), B cell translocation

gene 2 (BTG2) and protein arginine methyl transferase 1 (PRMT1) (Dolezal et al., 2017).

By comparison, we are only beginning to understand mechanisms that control

proliferation in pro-B cells. At this stage, cell cycle progression driven by IL-7 is

segregated from Igh recombination (Johnson et al., 2012) and the RNA-binding proteins

(RBPs) ZFP36L1 and ZFP36L2 suppress proliferation allowing Igh recombination and B

cell development (Galloway et al., 2016).

A network of transcription factors controls development and identity in B cells

(Busslinger, 2004). After transcription, RBPs control messenger RNA (mRNA) expression

and coordinate functionally related in mRNA regulons (Keene, 2007). Control of

3 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

cell cycle mRNA regulons by RBPs is fundamental for early lymphocyte development

(Galloway and Turner, 2017). Polypyrimidine Tract Binding Proteins (PTBP) are RBPs

that control , alternative polyadenylation, mRNA stability and

translation (Hu et al., 2018; Knoch et al., 2004; 2014). PTBP1 can either increase or

decrease mRNA stability by binding to the 3’UTR of transcripts or by modulating

alternative splicing of that will form transcript isoforms degraded by nonsense-

mediated mRNA decay (NMD) (Hu et al., 2018). PTBP1 is expressed ubiquitously. PTBP2

and PTBP3 (formerly ROD1) are highly conserved paralogs of PTBP1 expressed in

neurons and hematopoietic cells, respectively. PTBP1 suppresses expression of PTBP2

directly by inducing Ptbp2 10 skipping and mRNA degradation by NMD (Boutz et

al., 2007). In germinal centre B cells, PTBP1 promotes proliferation, the c-MYC gene

expression program induced upon positive selection and antibody affinity maturation

(Monzon-Casanova et al., 2018). PTBP1 and PTBP2 have specific and redundant functions

(Spellman et al., 2007; Vuong et al., 2016; Ling et al., 2016). Therefore, expression of

PTBP2 in PTBP1-deficient cells compensates for many functions of PTBP1 (Spellman et

al., 2007; Vuong et al., 2016; Ling et al., 2016; Monzon-Casanova et al., 2018).

To study the roles of PTBP1 controlling post-transcriptional in B cells,

we and others deleted Ptbp1 in pro-B cells and found normal B cell development

accompanied by PTBP2 expression (Monzon-Casanova et al., 2018; Sasanuma et al.,

2019). Here we addressed the functions of PTBP in B cell development by deleting both

Ptbp1 and Ptbp2 in pro-B cells. We show that PTBP1, and in its absence PTBP2, are

essential to promote B cell lymphopoiesis. In pro-B cells PTBP1 suppresses entry into S-

phase and promotes entry into mitosis after G2-phase. At the molecular level, PTBP1 is

4 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

necessary to maintain transcriptome fidelity in pro-B cells and is an essential component

of a previously unrecognised cell cycle mRNA regulon.

5 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Results

Redundancy between PTBP1 and PTBP2 for B cell development

As both PTBP1 and PTBP3 are expressed in mature B cells (Monzon-Casanova et al.,

2018) we determined their expression during defined stages of B cell development in

mouse bone marrow (Supplementary Fig. 1a-d). Fluorescence intensity of intracellular

PTBP1 staining by flow cytometry was greater in pro- and early-pre-B cells compared to

late-pre-B and immature-B cells. However, the isotype control staining for PTBP1 showed

the same differences in fluorescence intensity (Supplementary Fig. 1c, d). Therefore,

PTBP1 protein expression was similar between the different bone marrow B cell

populations. PTBP3 protein detected with a specific monoclonal antibody (Monzon-

Casanova et al., 2018) was also relatively constant throughout B cell development

(Supplementary Fig. 1c, d). PTBP2 protein was not detected in any of the B cell

developmental stages analysed (Supplementary Fig. 1c, d). Thus, PTBP1 and PTBP3, but

not PTBP2, are co-expressed throughout B cell development.

Conditional deletion of a Ptbp1-floxed allele (Suckale et al., 2011) in pro-B cells with the

Cd79acre allele (Cd79acre/+Ptbp1fl/fl mice, denoted here as Ptbp1 single conditional knock-

out, P1sKO) did not affect B cell development and resulted in the expression of PTBP2

(Monzon-Casanova et al., 2018). Therefore, we generated a double conditional knock-out

(dKO) mouse model where both Ptbp1 and Ptbp2 were deleted in pro-B cells using floxed

alleles and Cd79acre (Cd79acre/+Ptbp1fl/flPtbp2fl/fl mice, denoted here as P1P2dKO).

Deletion of both Ptbp1 and Ptbp2 in pro-B cells resulted in the absence of mature B cells

in the spleen and lymph nodes (Fig. 1a, b and Supplementary Fig. 1e).

6 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

PTBPs are required for progression beyond the pro-B cell stage

The absence of mature B cells in P1P2dKO mice resulted from a block in B cell

development at the pro-B cell stage (Fig. 1c, d and Supplementary Fig. 1a). We

confirmed the lack of PTBP proteins in pro-B cells from the conditional KO mice

(Supplementary Fig. 2a, b). PTBP1 protein was absent in P1sKO and P1P2dKO pro-B

cells and Ptbp1 deletion led to the expression of PTBP2 in P1sKO pro-B cells

(Supplementary Fig. 2b). P1P2dKO pro-B cells failed to increase PTBP2 in response to

Ptbp1 deletion, validating our dKO. P1P2dKO pro-B cells expressed increased amounts of

PTBP3 compared to pro-B cells from littermate control mice (Cd79a+/+Ptbp1fl/flPtbp2fl/fl,

denoted here as “control” unless stated otherwise) (Supplementary Fig. 2b). Yet, PTBP3

was insufficient to compensate for the lack of PTBP1 and PTBP2 in pro-B cells (Fig. 1c,

d), demonstrating an irreplaceable role for PTBP1 in B cell development in the absence

of PTBP2.

Characterisation of B cell development distinguishing between pro-, early- and late-pre-

B cells revealed an increase in the numbers of early-pre-B cells in P1sKO compared to

control mice and to Cd79acre “Cre-only” mice (mb1, Cd79aCre/+Ptbp1+/+Ptbp2+/+) (Fig. 1d).

In contrast, the numbers of late-pre- and immature-B cells were similar between P1sKO

and control mice (Fig. 1d). These data confirmed that deletion of Ptbp1 alone results in

largely unaffected B cell development.

In P1P2dKO mice there were similar numbers of FrB pro-B cells (B220+, CD19+, IgM-,

CD43high, CD24+, CD249low) and FrC pro-B cells (B220+, CD19+, IgM-, CD43high, CD24+,

CD249high) compared to littermate control and to Cd79acre “Cre-only” mice but we did not

find FrC’ early-pre-B cells, characterised by high expression of CD24 and CD249 (detected

7 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

by BP-1) (Fig. 1d). P1P2dKO FrB and FrC pro-B cells have higher CD24 staining than

control FrB and FrC pro-B cells (Fig. 1c). Therefore, we set different gates in the P1P2dKO

mice to include all these cells. FrB and FrC pro-B cells from P1P2dKO mice showed

increased c-KIT (CD117) staining compared to P1sKO and control pro-B cells and lacked

CD2, corroborating a block at an early developmental stage (Supplementary Fig. 2c, d).

With a cell surface and intracellular staining (Supplementary Fig. 2a) to identify early

and late pre-B cells as cells with successfully rearranged intracellular Igµ chain we

confirmed the block in B cell development in P1P2dKO mice at the pro-B cell stage

(Supplementary Fig. 2a, e). With this gating strategy the numbers of pro-B cells in

P1P2dKO mice were reduced compared to P1sKO and control mice. Few early-pre-B cells

(B220+, CD19+, IgM-, CD93+, CD43high, Igµ+) were present in P1P2dKO mice indicating

some successful recombination of the Igh locus in P1P2dKO pro-B cells (Supplementary

Fig. 2a). However, the numbers of early-pre-B cells were decreased by ~13-fold in

P1P2dKO compared to control mice (Supplementary Fig. 2e). Compared to the numbers

of pro-B cells in Rag2-/- deficient mice, the numbers of P1P2dKO pro-B cells identified by

the two different staining strategies were decreased by ~4.6-fold (FrB), 19-fold (FrC)

(Fig. 1d) and 47-fold (pro-B, Supplementary Fig. 2e). Thus, when Igh recombination is

defective in Rag2-/- mice developing B cells accumulate at the pro-B cell stage, whereas in

P1P2dKO mice pro-B cells do not accumulate.

PTBP1 suppresses entry into S-phase in pro-B cells

Dynamic control of proliferation is essential for successful B cell development (Clark et

al., 2014; Galloway and Turner, 2017). Therefore, we assessed the proliferative status of

pro-B cells in the P1P2dKO mice. To identify cells in early-, late- or post-S-phase we

sequentially labelled cells in vivo with EdU and BrdU (Fig. 2a). With this approach, cells

8 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

that incorporate only EdU were in S-phase at the beginning, but not at the end of the

labelling period and are in post S-phase; cells that incorporate both EdU and BrdU are in

late S-phase; and cells labelled only with BrdU at the end of the labelling period are in

early S-phase. In control mice, we detected few pre-pro-B cells in S-phase (3.8% in early-

and late-S together), compared to 13% of pro-B cells and high frequencies (53%) of early-

pre-B cells in S-phase (Supplementary Fig. 3a-d). In P1P2dKO pro-B cells the

proportions of early-S-phase cells were increased 4.7-fold compared to control and 3.2-

fold compared to P1sKO pro-B cells (Fig. 2b, c). The numbers of early-S-phase P1P2dKO

pro-B cells were similar to those of P1sKO and control pro-B cells (Fig. 2d). However, the

numbers of the whole pro-B cell population identified with the combination of cell

surface and intracellular markers (IgD-IgM-B220+CD19+CD43highIgµ+) used with the EdU

and BrdU labelling approach (Supplementary Fig. 3a) were reduced 4.4-fold in

P1P2dKO compared to control mice (Fig. 2d). Therefore, it is unlikely that the increased

proportions in early S-phase amongst P1P2dKO pro-B cells result from reduced numbers

of pro-B cells in G0/G1-phase. Thus, in the absence of PTBP1 and PTBP2 pro-B cells enter

S-phase more readily. The proportions of P1P2dKO pro-B cells in late S-phase and post

S-phase were also increased compared to P1sKO pro-B cells and control pro-B cells (Fig.

2c). The magnitude of the increase in the proportions of cells in post S-phase was smaller

than the increase in either early- or late-S-phase. This indicated that P1P2dKO pro-B cells

enter S-phase more readily than control pro-B cells but they fail to progress through S-

phase at the same pace as PTBP1-sufficient pro-B cells.

PTBP1 is required for the transition from G2 to M phase in pro-B cells

We also analysed cells in G2/M-phases by detecting DNA synthesis in combination with

DNA content (Fig. 2e, Supplementary Fig. 3e, f). The proportion of P1P2dKO pro-B cells

9 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

in G2/M-phase was increased ~40-fold compared to P1sKO and control pro-B cells (Fig.

2e, f) and the numbers of G2/M-phase pro-B cells in P1P2dKO mice were increased ~8-

fold compared to control and P1sKO mice (Fig. 2e, f). These findings were confirmed

when identifying FrB pro-B cells using cell surface markers (CD19+, CD24+, CD249-, CD2-

, CD25-, IgM-, IgD-), EdU labelling and DNA staining (Supplementary Fig. 3g, h). To

preserve phosphorylated histone 3 (pH3), which marks cells in mitosis, during the

staining process and quantitating DNA content with sufficient resolution, we identified

FrB pro-B cells again with cell surface markers (CD19+, CD24+, CD249- , CD2-, CD25-, IgM-

, IgD-) (Supplementary Fig. 3g, i). Some P1P2dKO pro-B cells had high pH3 levels

amongst cells with 4N-DNA content (Supplementary Fig. 3i). However, the proportions

of pH3-positive cells amongst P1P2dKO pro-B cells with 4N-DNA content were reduced

compared to the proportions found amongst P1sKO and control pro-B cells

(Supplementary Fig. 3i, j). Therefore, the majority of P1P2dKO pro-B cells with 4N-DNA

content are in G2-phase and have not entered mitosis. PTBP1 is thus required in pro-B

cells to progress from G2 to M-phase of the cell cycle.

A block in G2-phase is a hallmark response to DNA damage. Therefore, we assessed

markers of DNA damage at different stages of the cell cycle by flow cytometry

(Supplementary Fig. 4). G0/G1 P1P2dKO pro-B cells are bigger compared to P1sKO and

control G0/G1 pro-B cells (FSC-A measurements, Supplementary Fig. 4a). This is

consistent with a predisposition to enter S-phase, as size increase and S-phase entry are

positively correlated (Ginzberg et al., 2015). This increase in size resulted in higher

intracellular background staining with isotype-control antibodies in G0/G1 P1P2dKO

pro-B cells compared to G0/G1 P1sKO and control pro-B cells (Supplementary Fig. 4b,

c). pH2AX staining in P1P2dKO pro-B cells was not increased compared to P1sKO and

10 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

control pro-B cells above the increase seen in background staining controls

(Supplementary Fig. 4b, c). Staining of bone marrow cells treated with etoposide to

induce DNA damage showed that P1P2dKO pro-B cells increased pH2AX levels to the

same extent as P1sKO and control pro-B cells (Supplementary Fig. 4d, e). We found an

increase (2.1-fold) in p53 staining in P1P2dKO G0/G1 pro-B cells compared to control

G0/G1 pro-B cells above the increase (1.6-fold) seen in background staining controls

(Supplementary Fig. 4f, g). Upon etoposide treatment, the fold increase in p53 in

P1P2dKO G0/G1 pro-B cells compared to control pro-B cells (2.5-fold) was higher than

in pro-B cells without etoposide treatment (2.1-fold) (Supplementary Fig. 4g). Thus,

p53 levels provide some evidence of DNA damage in pro-B cells lacking PTBP1 and

PTBP2 but these DNA damage levels are insufficient to be detected by pH2AX-staining.

Therefore, DNA damage in P1P2dKO pro-B cells could contribute to the block in G2-

phase.

PTBP1 is necessary to maintain transcriptome fidelity in pro-B cells

To identify the consequences of PTBP deficiency on the transcriptome of pro-B cells we

carried out mRNAseq with polyadenylated transcripts from P1P2dKO, P1sKO and control

cKIT+ pro-B (FrB) cells (Supplementary Fig. 5). To identify candidate transcripts

directly bound and regulated by PTBP1 we made use of a PTBP1 individual-nucleotide

resolution Cross-Linking and ImmunoPrecipitation (iCLIP) (König et al., 2010) dataset

which reveals PTBP1-binding sites in the whole transcriptome and thereby allows

distinction between direct and indirect targets. The PTBP1 iCLIP data was from mitogen-

activated mouse primary B cells (Monzon-Casanova et al., 2018) because it was not

feasible to purify sufficient numbers of pro-B cells for iCLIP. There is a positive

correlation between the transcriptomes from mitogen-activated primary B cells and pro-

11 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

B cells (Supplementary Fig. 6a), suggesting that the available PTBP1 iCLIP data set

(Supplementary Tables 1, 2) allows to infer direct interactions of PTBP1 in pro-B cells.

We analysed changes in mRNA abundance by comparing pro-B cell transcriptomes from

the different genotypes in pairwise comparisons (Fig. 3a). There were more genes

showing differential mRNA abundance when comparing P1P2dKO to control pro-B cells

(1226 genes) and P1P2dKO to P1sKO pro-B cells (1202 genes) than when comparing

P1sKO to control pro-B cells (214 genes) (using a log2-transformed fold change >|0.5| and

FDR <0.05, Supplementary Table 3). We found that only ~25% of the genes with either

increased or decreased mRNA abundance due to absence of both PTBP1 and PTBP2 were

directly bound by PTBP1 at the 3’UTR and are therefore likely to be direct targets

(Supplementary Fig. 6b and Supplementary Table 3). The remaining changes

observed in mRNA abundance upon Ptbp1 and Ptbp2 deletion can be expected to be

independent from a direct role of PTBP1 via 3’UTR binding. Ptbp2 mRNA abundance was

increased ~18-fold in P1sKO compared to control pro-B cells (Fig. 3a, Supplementary

Table 3). Similarly, P1P2dKO pro-B cells showed an upregulation (~6.8-fold) of Ptbp2

mRNA compared to control pro-B cells (Fig. 3a, middle plot). However, Ptbp2 transcripts

in P1P2dKO pro-B cells lack Ptbp2 exon 4 due to the conditional knock-out

(Supplementary Fig. 6c), and hence do not result in PTBP2 protein expression

(Supplementary Fig. 2b).

We assessed alternative splicing (AS) differences in pro-B cells with different genotypes

by computing differences in exon inclusion levels in pairwise comparisons with rMATS

(Shen et al., 2014) (Fig. 3b, Supplementary Table 4). rMATS considers five different

types of alternatively spliced events: skipped exons, mutually exclusive exons, alternative

12 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

5’ and 3’ splice sites (SS) and retained (Supplementary Fig. 6d). The inclusion

level of a particular AS event is the proportion (scaled to 1) of transcripts containing the

alternatively spliced mRNA segment. We chose a threshold of an absolute inclusion level

difference >0.1 to identify substantial changes in AS. Similar to our observations in mRNA

abundance, the absence of both PTBP1 and PTBP2 in pro-B cells resulted in more

differentially alternatively spliced events (P1P2dKO – Ctrl, 1599 events in 1058 genes)

than the absence of PTBP1 (P1sKO – Ctrl, 865 events in 659 genes) (Fig. 3b). Depending

on the type of AS event analysed, from 30 to 50% of the events with changes in AS, when

comparing P1P2dKO to control pro-B cells, were bound in their vicinity by PTBP1

(Supplementary Fig. 6e, Supplementary Table 4). This implicates PTBP1 in controlling

these AS events directly. There is a small overlap of genes with changes in mRNA

abundance and also AS in the different pairwise comparisons (Fig. 3c), indicating that the

inclusion of certain exons may generate NMD targets. Such AS events promoting NMD are

most likely underestimated, since the RNA containing these events will be degraded and

it might be difficult to detect by mRNAseq.

As B cell development is largely unaffected in P1sKO mice we focused on differences in

mRNA abundance and AS unique to P1P2dKO pro-B cells compared to both P1sKO and

control pro-B cells to identify changes causing the developmental block in P1P2dKO mice.

We identified 1021 and 611 genes with decreased and increased mRNA abundance,

respectively, in P1P2dKO compared to control and P1sKO pro-B cells (Fig. 3d). Amongst

AS changes, we found 971 genes with increased or decreased inclusion levels of at least

one event in P1P2dKO pro-B cells compared to control and P1sKO (Fig. 3e). When

addressing abundance changes, P1P2dKO pro-B cells cluster separately from P1sKO and

control pro-B cells (Fig. 3d). In contrast, when analysing AS, P1P2dKO and P1sKO pro-B

13 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

cells cluster together and separately from control pro-B cells (Fig. 3e). Therefore, PTBP2

is more able to compensate for the absence of PTBP1 at the mRNA abundance than at the

AS level. We observed only few genes with different mRNA abundance (cluster d) or

changes in AS (clusters i and iii) that were predominantly regulated in P1sKO pro-B cells

compared to P1P2dKO and control pro-B cells (Fig. 3d, e). In contrast, large numbers of

genes had different mRNA abundance (clusters a, b and c) and AS (clusters ii and iv) in

P1P2dKO pro-B cells compared to P1sKO and control pro-B cells (Fig. 3d, e). Therefore,

in pro-B cells, PTBP2 has only few specific targets and mostly compensates for the

absence of PTBP1. We conclude that in pro-B cells PTBP1 ensures appropriate expression

at the level of AS and mRNA abundance level of more than 2000 genes and therefore

promotes the fidelity of their transcriptome.

PTBP1 controls CDK activity in pro-B cells

To assess functions of the genes with altered expression in P1P2dKO pro-B cells we

carried out (GO) enrichment analysis on genes with increased or reduced

abundance or with AS changes. We found numerous enriched GO terms important for the

biology of B cells and their progenitors (Fig. 4a, Supplementary Table 5) such as

“antigen processing and presentation of peptide antigen via MHC class II” amongst genes

with increased abundance and “positive regulation of DNA-binding transcription factor

activity” amongst genes with decreased abundance. We found enrichment of terms

related to migration such as “neuronal crest migration” and “neutrophil migration”. At

the AS level, enrichment of the term “cytoskeletal protein binding” suggests PTBP1 as a

cytoskeleton regulator. “Regulation of neuron differentiation” was enriched amongst

genes with reduced abundance and with changes in AS, as we expected from the known

roles of PTBPs in neuronal development (Hu et al., 2018). Thus, PTBP1 is necessary in

14 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

pro-B cells to promote adequate expression of many different pathways fundamental for

cell biology.

In P1P2dKO pro-B cells there were no obvious expression defects in genes important for

the progression from pro- to pre-B cells (Supplementary Fig. 6f, Supplementary

Tables 3, 4). Therefore, we focused on understanding the molecular mechanisms driving

uncontrolled entry into S-phase and increased frequencies of pro-B cells in G2-phase. We

only found one GO term amongst those enriched to be directly associated with

proliferation: “cyclin-dependent protein serine/threonine kinase inhibitor activity”. This

was unexpected given the documented changes in the transcriptome across different

phases of the cell cycle (Giotti et al., 2018) and that FrB pro-B cells from which we carried

out mRNAseq in P1P2dKO mice have a ~10% increase of cells in S-phase and ~12%

increase of cells in G2/M-phase compared to control and P1sKO FrB pro-B cells

(Supplementary Fig. 3g, h). Moreover, we did not find a global increase of S- and G2/M-

associated transcripts (Giotti et al., 2018) in the P1P2dKO pro-B transcriptome (Fig. 4b).

Instead, we found 8 genes with a reduction of mRNA abundance greater than 2-fold (<-1

Log2-fold change) in P1P2dKO compared to control pro-B cells (Supplementary Table

6). This indicates that the majority of the changes observed in the transcriptome of

P1P2dKO compared to control and P1sKO pro-B cells are not the result of comparing

populations with different proliferative status, but are reflective of an altered pro-B cell

transcriptome.

Amongst genes associated with “cyclin-dependent protein serine/threonine kinase

inhibitor activity”, we found reduced mRNA abundance in Cdkn1b, Cdkn1c, Cdkn2c and

Cdkn2d (encoding p27, p57, p18 and p19, respectively) due to PTBP1 and PTBP2 absence

15 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

(Fig. 5a). None of these CDK inhibitors has an obvious change in AS or a binding site for

PTBP1 in the 3’UTR (Supplementary Tables 2, 4). Therefore, the changes in mRNA

expression of CDK inhibitors are likely to be indirectly regulated by PTBP1. We stained

p27 protein across different phases of the cell cycle and found it to be reduced ~1.7-fold

in G0/G1 P1P2dKO pro-B cells compared to G0/G1 P1sKO and control pro-B cells (Fig.

5b, c), consistent with the decreased mRNA abundance observed in mRNA-Seq (Fig. 5a).

In contrast, p27 staining was increased ~7-fold in G2 P1P2dKO pro-B cells compared to

P1sKO and control G2 pro-B cells (Fig. 5b, c). These differences in p27 staining seen

between G0/G1 and G2 phases are consistent with an indirect role of PTBP1 in

controlling p27 expression. In addition to p27 staining, we assessed CDK activity by

measuring the extent of RB phosphorylation, which is started in G1 by CDKs and

promotes the entry into S-phase (Otto and Sicinski, 2017). Almost all P1P2dKO G0/G1

pro-B cells had phosphorylated RB whereas only ~30% of P1sKO and control G0/G1 pro-

B cells had phosphorylated RB (Fig. 5d, e), showing that P1P2dKO pro-B cells in G0/G1-

phase have abnormally high CDK activity. In contrast, P1P2dKO pro-B cells in G2-phase

had much lower proportions of cells with highly phosphorylated RB (~30%) compared

to G2 P1sKO and control pro-B cells (~70%) (Fig. 5d, e). This suggests an impaired CDK

activity amongst G2 P1P2dKO pro-B cells, consistent with high p27 levels. The

phosphorylation of threonine-592 on SAMHD1 is mediated by CDK1 (Cribier et al., 2013).

Proportions of phospho-T592-SAMHD1-positive cells amongst P1P2dKO G2 pro-B cells

were reduced compared to P1sKO and control G2 pro-B cells (Fig. 5d, e). These results

indicate a requirement of PTBPs to achieve the adequate CDK1 activity necessary to enter

mitosis. Thus, PTBPs are essential for controlling activity of CDKs in different phases of

the cell cycle in pro-B cells: to limit entry into S-phase and to promote progression

through mitosis.

16 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

PTBP1 controls expression of a cell cycle mRNA regulon in pro-B cells

The changes in expression of CDK inhibitors seemed an indirect consequence of Ptbp1

and Ptbp2 deletion and cell cycle gene sets were not enriched amongst genes with altered

mRNA expression in the P1P2dKO pro-B cells. Thus, instead of gene sets, it was possible

that individual genes with rate limiting properties for cell cycle progression, which could

also control the expression of CDK-inhibitors and CDK activities, were directly regulated

by PTBP1. Therefore, we considered further genes affected in the absence of PTBP1 and

PTBP2 but not in the absence of PTBP1 alone that are known to regulate the cell cycle

(Fig. 6). We identified nine candidate targets with direct PTBP1 binding and changes at

the levels of mRNA abundance and AS that constitute an mRNA regulon controlling the

cell cycle.

We found three genes (Foxo1, Btg2 and Rbl1) whose proteins inhibit S-phase entry with

reduced mRNA abundance in P1P2dKO compared to control and P1sKO pro-B cells (Fig.

6a). Additionally, two genes whose proteins promote entry into S-phase (Myc and Ccnd2)

showed increased mRNA abundance due to Ptbp1 and Ptbp2 deletion (Fig. 6a). FOXO

transcription factors promote p27 expression and inhibit CDK activity (Herzog et al.,

2009). PTBP1 bound to the 3’UTR of Foxo1 (Supplementary Fig. 7a). No obvious change

in Foxo1 AS was detected upon Ptbp1 and Ptbp2 deletion (Supplementary Table 4).

Therefore, PTBP1 expression enhances Foxo1 mRNA abundance in pro-B cells. PTBP1

bound to Myc 3’UTR (Supplementary Fig. 7a). We confirmed an increase of c-MYC

protein in G0/G1 P1P2dKO pro-B cells compared to control and P1sKO G0/G1 pro-B cells

(Supplementary Fig. 7b, c) as the fold-change increase in fluorescence intensity of the

c-MYC staining (2.2-fold) was greater than the fold-change increase in fluorescence

17 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

intensity of the control staining (1.5-fold). Therefore, in pro-B cells PTBP1 suppresses

Myc. CYCLIN-D2 pairs with CDK4/6. Ccnd2 abundance (encoding CYCLIN-D2) was

increased ~2-fold in P1P2dKO compared to control and P1sKO pro-B cells (Fig. 6a) and

PTBP1 bound to Ccnd2 3’UTR (Supplementary Fig. 7a). c-MYC induces transcription of

Ccnd2 (Bouchard et al., 2001). Therefore, PTBP1 and PTBP2 may reduce Ccnd2

abundance indirectly by suppressing c-MYC-mediated transcription and directly by

binding its mRNA. BTG2 inhibits G1 to S transition and promotes B cell development by

suppressing proliferation in late-pre-B cells (Dolezal et al., 2017). Btg2 mRNA abundance

was reduced ~2-fold due to Ptbp1 and Ptbp2 deletion (Fig. 6a). No obvious change in

Btg2 AS was detected but PTBP1 binding sites were present at the Btg2 3’UTR

(Supplementary Fig. 7a). Therefore, PTBP1 could directly promote BTG2 expression in

pro-B cells by stabilising Btg2 transcripts. p107 (encoded by Rbl1) represses E2F

transcription factors and entry into S-phase (Bertoli et al., 2013). Rbl1 mRNA abundance

was reduced ~1.5-fold in P1P2dKO compared to control and P1sKO pro-B cells (Fig. 6a).

PTBP1 and PTBP2 suppress the inclusion of a downstream alternative 5’SS (A5SS) in Rbl1

exon 8 which generates an NMD-target isoform in P1P2dKO pro-B cells (Fig. 6b). PTBP1

binds to adjacent regions of this alternatively spliced event (Fig. 6b). This A5SS event is

conserved in human, as PTBP1- and PTBP2-depleted HeLa cells (Ling et al., 2016) also

have an increased usage of the downstream A5SS compared to PTBP-sufficient cells

resulting in reduced RBL1 mRNA abundance (Fig. 6c, d). Therefore, PTBP1 promotes

Rbl1 expression. Collectively, deregulated expression of Foxo1, Myc, Ccnd2, Btg2 and Rbl1

in the absence of PTBP1 and PTBP2 could drive entry of pro-B cells into S-phase.

We also found transcripts important for the transition from G2 to M-phase altered in

P1P2dKO pro-B cells. The abundance of Cdc25b, Ect2, Kif2c and Kif22 was reduced (~3,

18 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

~5, ~2 and ~2-fold, respectively) in P1P2dKO compared to control and P1sKO pro-B cells

(Fig. 6a). CDC25B is a phosphatase promoting G2/M transition by activating CDK1

(Boutros et al., 2007). Cdc25b AS was unaffected by PTBP1 and PTBP2 absence and

PTBP1 bound to the Cdc25b 3’UTR (Supplementary Table 4 and Supplementary Fig.

7a). Thus, PTBP1 likely promotes Cdc25b expression by binding to its 3’UTR and

enhancing its stability. Additionally, there were three genes where their changes in

mRNA abundance can be explained by AS events leading to NMD. ECT2 controls spindle

formation in mitosis by exchanging GDP for GTP in small GTPases such as RhoA (Yüce et

al., 2005). In P1P2dKO pro-B cells Ect2 has an increased inclusion of an alternative exon

compared to P1sKO and control pro-B cells (Fig. 6b). Inclusion of this AS exon, using

either of two alternative 3’ splice sites, generates NMD-predicted transcripts. PTBP1

binds nearby this NMD-triggering exon where it is predicted to suppress exon inclusion

(Llorian et al., 2010) promoting high Ect2 mRNA levels in pro-B cells. KIF2c (MCAK) and

KIF22 (KID) are two kinesin motor family members that transport cargo along

microtubules during mitosis (Cross and McAinsh, 2014). Both have increased inclusion

of an exon that generates predicted NMD-targets in P1P2dKO compared to P1sKO and

control pro-B cells and some evidence for PTBP1 binding in adjacent regions (Fig. 6b).

This data shows that PTBP1 promotes expression of genes important for mitosis by

controlling AS linked to NMD. The altered expression of these genes in the absence of

PTBP1 and PTBP2 is likely to contribute to the high proportions of G2 cells observed in

P1P2dKO pro-B cells. Our findings place PTBP1 as a regulator of the cell cycle in pro-B

cells. PTBP1 is essential to control appropriate expression of an mRNA regulon dictating

CDK activity, progression to S-phase and entry into mitosis. In the absence of PTBP1 and

its partially redundant paralog PTBP2, the molecular control of the cell cycle in primary

pro-B cells is altered and B cell development is halted (Fig. 6e).

19 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Discussion

Here we present an essential role for PTBP1 and PTBP2 in controlling B cell development

and the proliferative status of pro-B cells. PTBP1 and PTBP2 deficiency in pro-B cells

triggered entry into S-phase and accumulation in G2-phase, causing a complete block in

B cell development. PTBP1 suppression of entry into S-phase was unanticipated since in

other systems, including germinal centre (GC) B cells, PTBP1 promoted proliferation

(Suckale et al., 2011; Shibayama et al., 2009; La Porta et al., 2016) and progression

through late S-phase (Monzon-Casanova et al., 2018). GC B cells did not tolerate deletion

of both PTBP1 and PTBP2 (Monzon-Casanova et al., 2018) and previous studies

addressing proliferation in PTBP1-deficient cells were done in the presence of PTBP2

(Suckale et al., 2011; Shibayama et al., 2009; La Porta et al., 2016). Therefore, the role of

PTBP1 in suppressing S-phase entry could be unique to pro-B cells or could be found in

other cell types if they resist the absence of PTBP1 and PTBP2 long enough to assess cell-

cycle progression.

Suppression of proliferation in pre-B cells is necessary to allow Igl recombination (Clark

et al., 2014). Similarly, in pro-B cells entry into S-phase will inhibit VDJ Igh recombination.

We have previously shown that the RNA-binding proteins ZFP36L1 and ZFP36L2

suppress proliferation in pro-B cells, facilitate Igh recombination and consequently

promote B cell development (Galloway et al., 2016). In the absence of PTBP1 and PTBP2,

expression of Zfp36l1 was normal and Zfp36l2 was increased ~1.5-fold (Supplementary

Table 3), but still there was no suppression of S-phase entry. This suggests the presence

of distinct mRNA regulons controlled by the different RBPs important to control cell cycle

20 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

entry. Our results reveal the importance of suppressing proliferation also in pro-B cells

for successful B cell development.

PTBP1 maintains the fidelity of the transcriptome in pro-B cells since the absence of

PTBP1 and PTBP2 caused changes in mRNA expression of 2000 genes. In particular,

PTBP1 expression controls a cell cycle mRNA regulon comprising Foxo1, Myc, Ccnd2,

Btg2, Rbl1, Cdc25b, Ect2, Kif22 and Kif2c. Proteins from this mRNA regulon control each

other and have downstream effects on other cell cycle regulators. FOXO proteins

suppress D-type cyclins (Schmidt et al., 2002) and promote p27 (Herzog et al., 2009). c-

MYC suppresses Cdkn1b (encoding p27) (Yang et al., 2001) and increases Ccnd2

transcription (Bouchard et al., 2001). CYCLIN-D2 promotes nuclear export and

degradation of p27 (Susaki et al., 2007) and p27 can directly inhibit CYCLIN-D-CDK4/6

complexes (Yoon et al., 2012). We found evidence for interaction of PTBP1 with these

transcripts implicating a direct role of PTBP1. Future investigations are needed to assess

which of the PTBP1-binding sites are functional and control mRNA expression. However,

since these targets regulate each other it will be difficult to deconvolute direct from

indirect effects of PTBP1 controlling this cell cycle mRNA regulon. The net outcome from

deregulation of these transcripts in the absence of PTBP1 and PTBP2 is an enhanced CDK

activity in G0/G1 pro-B cells that drives entry into S-phase and a reduced CDK activity

amongst blocked G2 pro-B cells.

Myc mRNA and protein abundance was increased due to Ptbp1 and Ptbp2 deletion in pro-

B cells and PTBP1 was bound to the Myc 3’UTR. In GC B cells lacking PTBP1 alone, part of

the c-MYC-activated gene expression program initiated upon positive selection was

reduced while c-MYC protein expression was unaffected (Monzon-Casanova et al., 2018).

21 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

PTBP2 is redundant with PTBP1 in GC B cells (Monzon-Casanova et al., 2018), therefore

it is possible that c-MYC protein would be increased in GC B cells if both PTBP1 and PTBP2

are absent. We were unable to address this because we did not detect GC B cells lacking

both PTBP1 and PTBP2 (Monzon-Casanova et al., 2018). Alternatively, PTBP1 could have

distinct roles in controlling c-MYC and its gene expression programme in different cell

types.

Deletion of Ptbp1 and Ptbp2 in pro-B cells resulted in a block in G2-phase. Since the

number of pro-B cells did not accumulate in P1P2dKO mice, it is likely that P1P2dKO pro-

B cells die after the block at G2-phase. We found increased p53 in pro-B cells deficient for

PTBP1 and PTBP2, indicating that DNA damage in pro-B cells contributes to the G2 block,

although we did not detect a DNA damage signature amongst GO enrichment analysis.

DNA damage in P1P2dKO pro-B cells could happen due to the enhanced entry into S-

phase as this causes replication stress (Zeman and Cimprich, 2013). Moreover, p27

promotes G2 arrest in response to DNA-damage (Cuadrado et al., 2009; Payne et al.,

2008). Thus, the increased p27 levels found in P1P2dKO pro-B cells blocked at G2-phase

could be a response to DNA damage. Increased p27 will inhibit CDK activity and promote

G2 block (Yoon et al., 2012). PTBP1 promotes p27 IRES-mediated translation in human

cells (Cho et al., 2005). The sequence in the human CDKN1B (encoding p27) 5’UTR-IRES

is conserved in mouse. We did not find mouse PTBP1 bound to the Cdkn1b 5’UTR in our

iCLIP data. However, a direct role of PTBP1 in promoting p27 translation could still be

possible in pro-B cells and is consistent with a reduction in p27 in G0/G1 pro-B cells due

to PTBP1 and PTBP2 absence. Nevertheless, we found increased p27 levels in G2 pro-B

cells due to Ptbp1 and Ptbp2 deletion and this would be inconsistent with PTBP1

promoting p27 expression. Additionally, CDC25B reduction has been shown to cause a

22 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

G2 block (Peco et al., 2012; Lindqvist et al., 2005). Therefore, Cdc25b reduction in

P1P2dKO pro-B cells, in addition to the other mechanisms discussed, could also

contribute to the block in G2-phase.

Various RBPs promote transcriptome fidelity by suppressing non-conserved cryptic

exons which in normal conditions should never be included and often generate NMD

targets (Ling et al., 2016; Sibley et al., 2016). Cryptic exons suppressed by PTBP1 have

been found when both PTBP1 and PTBP2 were depleted (Ling et al., 2016). These cryptic

exons are poorly conserved between mouse and human. In pro-B cells deficient for

PTBP1 and PTBP2 we found cryptic exons promoting mRNA degradation amongst cell

cycle regulators. For most of them we did not find evidence for their conservation in

human cells where PTBP1 and PTBP2 were absent (Ling et al., 2016). However, the A5SS

in Rbl1 generating an NMD-target was conserved in human, suggesting that it is a

functional AS-NMD event with a role in post-transcriptional control. Altogether, PTBP1

expression in pro-B cells controls correct expression of thousands of genes. In particular,

PTBP1 controls a network of factors essential to avoid S-phase entry followed by a G2

block and to promote B cell development beyond the pro-B cell stage.

23 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Methods

Mice

Mice were bred and maintained in the Babraham Institute Biological Support Unit under

Specific Opportunistic Pathogen Free (SOPF) conditions. After weaning, mice were

transferred to individually ventilated cages with 1-5 mice per cage. All experiments were

carried out without blinding or randomization on 7 to 15-week-old mice. Conditional

knockout mice used were derived from crossing the following transgenic strains:

Ptbp1fl/fl(Ptbp1tm1Msol) (Suckale et al., 2011), Ptbp2fl/fl(Ptbp2tm1.1Dblk) (Li et al., 2014) and

Cd79acre(Cd79atm1(cre)Reth) (Hobeika et al., 2006). Rag2-/- knockout mice (Shinkai et al.,

1992) (Rag2tm1Fwa) were also used. All mice were on a C57BL/6 background.

In vivo EdU and BrdU administration

All procedures performed were approved by the Babraham Institute's Animal Welfare

and Experimentation Committee and the UK Home Office and followed all relevant ethical

regulations. In EdU-only and in EdU and BrdU double-labelling experiments 1 to 5 mice

of different genotypes and the same sex were kept in individually ventilated cages.

Whenever possible females were used as this allowed for a higher number of mice with

different genotypes per cage. Males and females showed the same differences in the

phenotypes observed due to PTBPs absence. In Edu-only labelling experiments mice

were injected with 1 mg EdU (5-ethynyl-2’-deoxyuridine, cat #E10415, ThermoFisher

Scientific) was injected intraperitoneally and mice were culled one hour after injection.

In EdU and BrdU (5-Bromo-2′-deoxyuridine, cat# B5002-500mg, Sigma) labelling

experiments mice were injected first with 1 mg EdU (cat #E10415, ThermoFisher

24 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Scientific) intraperitoneally, one hour later, the same mice were injected with 2 mg BrdU

and the mice were culled 1 hour and 40 minutes after the injection with EdU.

Flow cytometry

Single cell suspensions were prepared from spleens and lymph nodes by passing the

organs through cell strainers with 70 μm and 40 μm pore sizes in cold RPMI-1640 (cat#

R8758, Sigma) with 2% fetal calf serum (FCS). Single cell suspensions from bone

marrows were prepared by flushing the bone marrow from femurs and tibias and passing

the cells through a cell strainer with 40 μm pore size in cold RPMI-1640 with 2%FCS. Fc

receptors were blocked with monoclonal rat antibody 2.4G2. Cells were stained with

combinations of antibodies listed in Supplementary Table 7. Cell surface staining was

carried out for 45 minutes on ice by incubating cells with a mixture of antibodies in cold

FACS buffer (PBS +0.5%FCS). For intracellular staining cells were fixed with

Cytofix/Cytoperm™ Fixation and Permeabilization Solution (cat# 554722, BD) on ice,

washed with FACS Buffer and frozen in 10% DMSO 90% FCS at -80oC at least overnight.

After thawing, cells were washed with FACS Buffer and re-fixed with Cytofix/Cytoperm™

Fixation and Permeabilization Solution (cat# 554722, BD) for 5 minutes on ice. Cells were

washed with Perm/Wash Buffer (cat# 554723, BD) and intracellular staining with

antibodies were carried out by incubating first fixed and permeabilized cells in

Perm/Wash Buffer (cat# 554723, BD) with monoclonal rat antibody 2.4G2 and

subsequently with the desired antibodies in Perm/Wash Buffer at room temperature.

EdU was detected with Click-iT™ Plus EdU kits (cat# C10646, for AF594 and cat# C10636

for pacific blue, ThermoFisher Scientific). For double detection of EdU and BrdU, cells

were treated as for intracellular staining, but before adding the intracellular antibodies,

cells were treated with TURBO™ DNase (12 units/10^7 cells, cat# AM2239,

25 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

ThermoFisher Scientific) for 1 hour at 37oC, subsequently, EdU was detected with the

Click-iT™ reaction following the instructions from the manufacturer, cells were washed

with Perm/Wash Buffer (cat# 554723, BD) and incubated with anti-BrdU-AF647

antibody (MoBU-1, cat# B35133, ThermoFisher Scientific). DNA was stained in the last

step before flow cytometry analysis with 7AAD in EdU and BrdU double-labelling

experiments or with Vybrant™ DyeCycle™ Violet Stain in experiments where no DNA-

digestion was carried out (cat # V35003, ThermoFisher Scientific). In experiments where

cells were treated with etoposide to induce DNA damage, single cell suspensions from

bone marrow were incubated with 20μM etoposide (cat# E1383-100MG, Sigma) in RPMI

+2%FCS for 3 hours at 37oC. Flow cytometry data was acquired on a BD LSRFortessa with

5 lasers and was analysed using FlowJo software (versions 10.6.0 and 10.0.8r1).

mRNAseq libraries from FrB pro-B cells

FrB c-KIT+ pro-B cells (B220+, CD19+, IgD-, IgM-, CD2-, CD25-, CD43high, cKIT+, CD24+ and

CD249-) were sorted from bone marrow cells isolated from femurs and tibias. Bone

marrow cells from 4-6 mice from the same genotype and sex were pooled and depleted

of unwanted cells with anti Gr-1-bio (RB6-8C5), CD11b (M1/70), IgD-bio (11-26c.2a),

NK1.1-bio, CD3e (145-2C11) and Ter119 biotinylated antibodies and anti-biotin

microbeads (cat# 130-090-485, Miltenyi) before sorting. Sorting strategy of FrB pro-B

cells is shown in Supplementary Fig. 5. RNA from 15,000 to 200,000 FrB cells was

isolated with the RNeasy Micro Kit (cat# 74004, Qiagen). mRNAseq libraries from 5

biological replicates per genotype: (three genotypes: ctrl (Cd79a+/+Ptbp1fl/flPtbp2fl/fl),

P1sKO (Cd79acre/+Ptbp1fl/flPtbp2+/+) and P1P2dKO (Cd79acre/+Ptbp1fl/flPtbp2fl/fl); three

samples from females and two samples from males per genotype) were prepared by

generating cDNA from 2 ng RNA and 9 PCR cycles per replicate with the SMART-Seq® v4

26 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Ultra® Low Input RNA Kit for Sequencing (cat# 634891, Takara) and by enzymatic

fragmentation of 300 pg of cDNA followed by 12 PCR cycles using the Nextera XT DNA

Library Preparation Kit (cat# FC-131-1096, Illumina). Sequencing of the 15 mRNAseq

libraries multiplexed together was carried out with an Illumina HiSeq2500 on a 2x125

bp paired-end run.

PTBP1 iCLIP

PTBP1 iCLIP was carried out previously from mitogen-activated B cells (splenic B cells

stimulated with LPS for 48 hours)(Monzon-Casanova et al., 2018). This data was re-

analysed to map reads using a splicing-aware software. Reads from five PTBP1 iCLIP

libraries were mapped to the GRCm38.p5 mouse genome from Gencode with STAR (v

2.5.4b)(Dobin et al., 2013). Reads were de-duplicated using random barcodes included in

the library preparation and xlink-sites (Supplementary Table 1) and clusters of binding

(Supplementary Table 2) were identified with iCount

https://icount.readthedocs.io/en/latest/cite.html as previously described(Monzon-

Casanova et al., 2018). Detection of a binding site with iCLIP(König et al., 2010) is highly

dependent on the abundance of the RNA, therefore all replicates were pooled together to

identify xlink sites and clusters of binding. Xlink sites and clusters of PTBP1 binding were

assigned to transcripts and genomic features with the following hierarchy: CDS, 3’UTR,

5’UTR, , ncRNA using the Mus_musculus.GRCm38.93.gtf annotation from Ensembl

(Supplementary Tables 1, 2)

mRNA abundance analysis

mRNAseq libraries were trimmed with Trim Galore (v 1.15

https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) with default

27 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

parameters and mapped with Hisat2 (v 2.1.0) (Kim et al., 2015) with -p 7 -t --phred33-

quals --no-mixed --no-discordant parameters, the Mus_musculus.GRCm38 genome and

known splice sites from the Mus_musculus.GRCm38.90 annotation. Read counts mapped

over genes were counted with HTSeq (v0.10.0) (Anders et al., 2015) with f bam -r name

-s no -a 10 parameters and the Mus_musculus.GRCm38.93.gtf annotation from Ensembl.

Reads mapping to immunoglobulin genes (including V, D, J gene segments, light and

heavy immunoglobulin genes and T cell receptor genes) were excluded before DESeq2

analysis. Differences in mRNA abundance were computed with DESeq2 (v1.22.1) (Love

et al., 2014) by extracting differences between different genotypes in pair-wise

comparisons using “apeglm” method as a shrinkage estimator (Zhu et al., 2019).

Information on the sex of the mice from which the mRNAseq libraries were generated

was included in the design formula in addition to the genotype (design = ~ Sex +

Genotype). From the DESeq2 results we only considered genes with a mean expression

level of at least 1 FPKM (from the five biological replicates) in any of the three genotypes

analysed. FPKMs were calculated with cuffnorm from Cufflinks (v2.2.1) (Trapnell et al.,

2012) using the -library-norm-method geometric. We considered genes with differential

mRNA abundance as those with a log2-fold change >|0.5| and a p-adjusted value of <0.05

(Supplementary Table 3).

A gene with different mRNA abundance was bound by PTBP1 at the 3’UTR if at least one

cluster of PTBP1-binding from the PTBP1 iCLIP data (Supplementary Table 2) was

found to overlap with any 3’UTRs annotated in the Mus_musculus.GRCm38.93.gtf for that

gene after assignation to genomic features for the binding sites of the iCLIP as described

above.

28 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Genes with different mRNA abundance in P1P2dKO FrB pro-B cells compared to P1sKO

and control FrB cells (Fig. 3d, Supplementary Table 5) were identified after

hierarchical clustering of the Euclidian distances between Z-scores of each gene

calculated from the DESeq2 normalised read counts for the 15 mRNAseq libraries (5

biological replicates per genotype). Genes with different mRNA abundance in any of the

three pairwise comparisons carried out (Fig. 3a) were considered in the hierarchical

clustering.

Alternative splicing analysis

Trimmed reads with Trim Galore were further trimmed to 123 bp and reads which were

smaller than 123 bp were discarded with Trimmomatic (v0.35) (Bolger et al., 2014) in

order to obtain only pairs of reads 123 bp long. 123 bp-long reads were mapped to the

Mus_musculus.GRCm38 genome as described above, but only uniquely-mapped reads

were kept by using the -bS -F 4 -F 8 -F 256 -q 20 parameters in samtools when converting

hisat2 sam files to bam files. rMATS (Turbo v4.0.2) (Shen et al., 2014) was run with the -

t paired --readLength 123 parameters and the Mus_musculus.GRCm38.93.gtf annotation

for each individual pairwise comparison. Only results from rMATS with reads on exon-

exon junctions were considered further. Significantly differentially alternatively spliced

events (Supplementary Table 4) were defined as events that have an FDR<0.05, have

an absolute inclusion level difference >0.1, come from genes expressed with at least 1

FPKM (mean across five biological replicates in any of the genotypes analysed) and have

at least 80 reads from the sum of the five biological replicates mapping to either the

included or skipped alternatively spliced event in at least one of the two conditions

analysed. Only alternatively spliced events with the highest inclusion level difference

29 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

were kept from AS events with had the same genomic coordinates in the AS event

(Supplementary Table 4).

Proportions of differentially alternatively spliced events were defined as bound by PTBP1

(Supplementary Fig. 6e, Supplementary Table 4) with different rules depending on

the type of AS event. An SE was bound by PTBP1 if PTBP1 clusters (Supplementary

Table 2) were found on the SE, 500 nucleotides upstream or downstream of the AS SE,

on either of the constitutive flanking exons or in 500 nucleotides downstream or

upstream of the upstream and downstream constitutive exons, respectively. An A5SS was

bound if PTBP1 clusters were found on the longest exon containing the A5SS, on the

downstream intronic 500 nucleotides of the A5SS, on the downstream constitutive

flanking exon or the 500 intronic nucleotides upstream of the downstream constitute

flanking exon. An A3SS was bound by PTBP1 if PTBP1 clusters were found on the longest

exon containing the A3SS, the 500 intronic nucleotides upstream of the A3SS, the

upstream constitutive flanking exon or the 500 intronic nucleotides downstream of the

upstream constitutive flanking exon. A MXE was bound by PTBP1 if PTBP1 clusters were

found on either MXE, the upstream or downstream 500 intronic nucleotides of either

MXE, the upstream or downstream constitutive flanking exons or the downstream or

upstream 500 intronic nucleotides from the upstream or downstream constitutive

flanking exons, respectively. A RI was bound by PTBP1 if PTBP1 clusters were found on

the RI or on either constitutive flanking exon.

Events differentially alternatively spliced in P1P2dKO FrB pro-B cells compared to

control and P1sKO FrB pro-B cells (Fig. 3e) were identified by hierarchical clustering

using complete linkage clustering of the Euclidean distances between Z-scores of each AS

30 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

event calculated from the rMATS inclusion levels for the 15 mRNAseq libraries (5

biological replicates per genotype). AS events found in any of the three pairwise

comparisons (Fig. 3b) were used in the hierarchical clustering.

Gene ontology term enrichment analysis

Genes belonging to different clusters based on differences in mRNA abundance or AS

patterns between the P1P2dKO FrB pro B cells and the other genotypes (P1sKO and

Controls) (Fig. 3e, d and Supplementary Table 5) were used for gene ontology

enrichment analysis with GOrilla (Eden et al., 2009). Genes expressed with a mean of least

1 FPKM across the 5 biological replicates in any of the genotypes were used as

background list for expressed genes in FrB pro-B cells. Fig. 4a shows representative

enriched terms selected amongst closely related GO terms by manual inspection of the

ontology. Supplementary Table 5 is a full list of all enriched gene ontology enriched

terms. Representative selected terms are highlighted.

Comparison of transcriptomes from pro-B cells and mitogen activated B cells

Transcriptomes from control FrB pro-B cells and mitogen activated primary B cells (LPS

for 48 hours) (Diaz-Muñoz et al., 2015) were compared by calculating Spearman's rank

correlation of the Log mean TPM (Transcripts per million reads) for each gene. Mean

TPMs were calculated in FrB pro-B cells from five biological replicates and in mitogen-

activated B cells from four biological replicates. TPMs were calculated after counting

reads mapping to genes with HTSeq and the Mus_musculus.GRCm38.93.gtf annotation

from Ensembl.

mRNAseq from human PTBP1 and PTBP2 depleted HeLa

31 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Control and mRNAseq libraries where PTBP1 and PTBP2 were knocked down in HeLa

cells by Ling et al. (Ling et al., 2016) were trimmed with Trim Galore (v 0.6.2_dev) and

mapped with Hisat2 (v 2.1.0) (Kim et al., 2015) with --dta --sp 1000,1000 -p 7 -t --

phred33-quals --no-mixed --no-discordant parameters, the Homo_sapiens.GRCh38

genome annotation and a file with known splice sites generated from the

Homo_sapiens.GRCh38.87 annotation. Mapped reads were counted with HTSeq

(v0.10.0)(Anders et al., 2015) with -f bam -r name -s no -a 10 parameters and the

Homo_sapiens.GRCh38.93.gtf annotation from Ensembl. Read counts were normalised

with DESeq2 (Love et al., 2014) (v 1.20.0).

Statistical analysis

Statistical analysis of flow cytometry data was carried out with GraphPad Prism version

7.0e. Details of tests carried out are found in the legends. Statistical analysis of mRNAseq

data was carried out as described in the mRNA abundance analysis and Alternative

splicing analysis sections.

Data availability

mRNAseq libraries and iCLIP analysis generated in this study have been deposited in GEO

and can be accessed with the GSE136882 accession code at GEO. Mitogen activated

primary B cell mRNAseq libraries were previously reported and can be accessed with the

GSM1520115, GSM1520116, GSM1520117and GSM1520118 accession codes in GEO.

32 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Author contributions

E.M.C., M.T. and C.W.J.S. designed experiments. E.M.C. and K.T. performed experiments

and analysed data. E.MC., L.S.M. and K.Z. designed and carried out computational analysis.

E.M.C. and M.T. wrote the manuscript with input from the co-authors.

Acknowledgments

We thank Douglas L. Black for the Ptbp2fl/fl mice, Michael Reth for the Cd79acre mice,

Frederick W. Alt for the Rag2-/- mice and Michele Solimena for the Ptbp1fl/fl mice and the

anti-PTBP2 antibody; Kirsty Bates, Arthur Davis, Attila Bebes, Simon Andrews, the

Biological Support Unit, Flow Cytometry and Bioinformatics Facilities for technical

assistance; Tony Ly, Anne Corcoran and members of the Turner, Smith and Zarnack

laboratories for helpful discussions. This study was supported by funding from the

Biotechnology and Biological Sciences Research Council (BB/J00152X/1;

BB/P01898X/1; BBS/E/B/000C0407; and BBS/E/B/000C0427 and the Campus

Capability Core Grant to the Babraham Institute), a Wellcome Investigator award to M.T

and a Sort Term Scientific Mission Award to E.M.C. from the European Cooperation in

Science and Technology action CA17103, Delivery of Antisense RNA Therapeutics. The

authors declare no competing financial interests.

33 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

References

Anders, S., P.T. Pyl, and W. Huber. 2015. HTSeq--a Python framework to work with high- throughput sequencing data. Bioinformatics. 31:166–169. doi:10.1093/bioinformatics/btu638.

Bertoli, C., J.M. Skotheim, and R.A.M. de Bruin. 2013. Control of cell cycle transcription during G1 and S phases. Nature Reviews Molecular Cell Biology. 14:518–528. doi:10.1038/nrm3629.

Bolger, A.M., M. Lohse, and B. Usadel. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 30:2114–2120. doi:10.1093/bioinformatics/btu170.

Bouchard, C., O. Dittrich, A. Kiermaier, K. Dohmann, A. Menkel, M. Eilers, and B. Lüscher. 2001. Regulation of cyclin D2 gene expression by the Myc/Max/Mad network: Myc- dependent TRRAP recruitment and histone acetylation at the cyclin D2 promoter. Genes Dev. 15:2042–2047. doi:10.1101/gad.907901.

Boutros, R., V. Lobjois, and B. Ducommun. 2007. CDC25 phosphatases in cancer cells: key players? Good targets? Nature Reviews Cancer 2017 17:2. 7:495–507. doi:10.1038/nrc2169.

Boutz, P.L., P. Stoilov, Q. Li, C.-H. Lin, G. Chawla, K. Ostrow, L. Shiue, M. Ares, and D.L. Black. 2007. A post-transcriptional regulatory switch in polypyrimidine tract- binding proteins reprograms alternative splicing in developing neurons. Genes Dev. 21:1636–1652. doi:10.1101/gad.1558107.

Busslinger, M. 2004. Transcriptional Control of Early B Cell Development1. http://dx.doi.org/10.1146/annurev.immunol.22.012703.104807. 22:55–79. doi:10.1146/annurev.immunol.22.012703.104807.

Cho, S., J.H. Kim, S.H. Back, and S.K. Jang. 2005. Polypyrimidine Tract-Binding Protein Enhances the Internal Ribosomal Entry Site-Dependent Translation of p27Kip1 mRNA and Modulates Transition from G1 to S Phase. Mol. Cell. Biol. 25:1283–1297. doi:10.1128/MCB.25.4.1283-1297.2005.

Clark, M.R., M. Mandal, K. Ochiai, and H. Singh. 2014. Orchestrating B cell lymphopoiesis through interplay of IL-7 receptor and pre-B cell receptor signalling. Nature Reviews Immunology. 14:69–80. doi:10.1038/nri3570.

Cribier, A., B. Descours, A.L.C. Valadão, N. Laguette, and M. Benkirane. 2013. Phosphorylation of SAMHD1 by Cyclin A2/CDK1 Regulates Its Restriction Activity toward HIV-1. Cell Reports. 3:1036–1043. doi:10.1016/j.celrep.2013.03.017.

Cross, R.A., and A. McAinsh. 2014. Prime movers: the mechanochemistry of mitotic kinesins. Nature Reviews Molecular Cell Biology. 15:257–271. doi:10.1038/nrm3768.

Cuadrado, M., P. Gutierrez-Martinez, A. Swat, A.R. Nebreda, and O. Fernandez-Capetillo. 2009. p27Kip1 Stabilization Is Essential for the Maintenance of Cell Cycle Arrest in

34 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Response to DNA Damage. Cancer Res. 69:8726–8732. doi:10.1158/0008- 5472.CAN-09-0729.

Diaz-Muñoz, M.D., S.E. Bell, K. Fairfax, E. Monzon-Casanova, A.F. Cunningham, M. Gonzalez-Porta, S.R. Andrews, V.I. Bunik, K. Zarnack, T. Curk, W.A. Heggermont, S. Heymans, G.E. Gibson, D.L. Kontoyiannis, J. Ule, and M. Turner. 2015. The RNA- binding protein HuR is essential for the B cell antibody response. Nat. Immunol. 16:415–425. doi:10.1038/ni.3115.

Dobin, A., C.A. Davis, F. Schlesinger, J. Drenkow, C. Zaleski, S. Jha, P. Batut, M. Chaisson, and T.R. Gingeras. 2013. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 29:15–21. doi:10.1093/bioinformatics/bts635.

Dolezal, E., S. Infantino, F. Drepper, T. Börsig, A. Singh, T. Wossning, G.J. Fiala, S. Minguet, B. Warscheid, D.M. Tarlinton, H. Jumaa, D. Medgyesi, and M. Reth. 2017. The BTG2- PRMT1 module limits pre-B cell expansion by regulating the CDK4-Cyclin-D3 complex. Nat. Immunol. 18:911–920. doi:10.1038/ni.3774.

Eden, E., R. Navon, I. Steinfeld, D. Lipson, and Z. Yakhini. 2009. GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics. 10:48. doi:10.1186/1471-2105-10-48.

Galloway, A., A. Saveliev, S. Łukasiak, D.J. Hodson, D. Bolland, K. Balmanno, H. Ahlfors, E. Monzon-Casanova, S.C. Mannurita, L.S. Bell, S. Andrews, M.D. Diaz-Muñoz, S.J. Cook, A. Corcoran, and M. Turner. 2016. RNA-binding proteins ZFP36L1 and ZFP36L2 promote cell quiescence. Science. 352:453–459. doi:10.1126/science.aad5978.

Galloway, A., and M. Turner. 2017. Cell cycle RNA regulons coordinating early lymphocyte development. Wiley Interdisciplinary Reviews: RNA. 8:e1419. doi:10.1002/wrna.1419.

Ginzberg, M.B., R. Kafri, and M. Kirschner. 2015. On being the right (cell) size. Science. 348:1245075–1245075. doi:10.1126/science.1245075.

Giotti, B., S.-H. Chen, M.W. Barnett, T. Regan, T. Ly, S. Wiemann, D.A. Hume, and T.C. Freeman. 2018. Assembly of a Parts List of the Human Mitotic Cell Cycle Machinery. J Mol Cell Biol. 32:2858. doi:10.1093/jmcb/mjy063.

Hardy, R.R., and K. Hayakawa. 2003. B Cell Development Pathways. https://doi.org/10.1146/annurev.immunol.19.1.595. 19:595–621. doi:10.1146/annurev.immunol.19.1.595.

Hardy, R.R., C.E. Carmack, S.A. Shinton, J.D. Kemp, and K. Hayakawa. 1991. Resolution and characterization of pro-B and pre-pro-B cell stages in normal mouse bone marrow. Journal of Experimental Medicine. 173:1213–1225. doi:10.1084/jem.173.5.1213.

Herzog, S., M. Reth, and H. Jumaa. 2009. Regulation of B-cell proliferation and differentiation by pre-B-cell receptor signalling. Nature Reviews Immunology. 9:195–205. doi:10.1038/nri2491.

35 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Hobeika, E., S. Thiemann, B. Storch, H. Jumaa, P.J. Nielsen, R. Pelanda, and M. Reth. 2006. Testing gene function early in the B cell lineage in mb1-cre mice. PNAS. 103:13789– 13794. doi:10.1073/pnas.0605944103.

Hu, J., H. Qian, Y. Xue, and X.-D. Fu. 2018. PTB/nPTB: master regulators of neuronal fate in mammals. Biophys Rep. 4:204–214. doi:10.1007/s41048-018-0066-y.

Johnson, K., J. Chaumeil, M. Micsinai, J.M.H. Wang, L.B. Ramsey, G.V. Baracho, R.C. Rickert, F. Strino, Y. Kluger, M.A. Farrar, and J.A. Skok. 2012. IL-7 Functionally Segregates the Pro-B Cell Stage by Regulating Transcription of Recombination Mediators across Cell Cycle. The Journal of Immunology. 188:6084–6092. doi:10.4049/jimmunol.1200368.

Keene, J.D. 2007. RNA regulons: coordination of post-transcriptional events. Nature Reviews Genetics 2013 14:7. 8:533–543. doi:10.1038/nrg2111.

Kim, D., Ben Langmead, and S.L. Salzberg. 2015. HISAT: a fast spliced aligner with low memory requirements. Nat. Methods. 12:357–360. doi:10.1038/nmeth.3317.

Knoch, K.-P., H. Bergert, B. Borgonovo, H.-D. Saeger, A. Altkrüger, P. Verkade, and M. Solimena. 2004. Polypyrimidine tract-binding protein promotes insulin secretory granule biogenesis. Nature Cell Biology. 6:207–214. doi:10.1038/ncb1099.

Knoch, K.-P., S. Nath-Sain, A. Petzold, H. Schneider, M. Beck, C. Wegbrod, A. S nmez, C. M nster, A. Friedrich, M. Roivainen, and M. Solimena. 2014. PTBP1 is required for glucose-stimulated cap-independent translation of insulin granule proteins and Coxsackieviruses in beta cells. Molecular Metabolism. 3:518–530.

König, J., K. Zarnack, G. Rot, T. Curk, M. Kayikci, B. Zupan, D.J. Turner, N.M. Luscombe, and J. Ule. 2010. iCLIP reveals the function of hnRNP particles in splicing at individual nucleotide resolution. Nat. Struct. Mol. Biol. 17:909–915. doi:10.1038/nsmb.1838.

La Porta, J., R. Matus-Nicodemos, A. Valentín-Acevedo, and L.R. Covey. 2016. The RNA- Binding Protein, Polypyrimidine Tract-Binding Protein 1 (PTBP1) Is a Key Regulator of CD4 T Cell Activation. PLoS ONE. 11:e0158708. doi:10.1371/journal.pone.0158708.

Li, Q., S. Zheng, A. Han, C.-H. Lin, P. Stoilov, X.-D. Fu, D.L. Black, and B.J. Blencowe. 2014. The splicing regulator PTBP2 controls a program of embryonic splicing required for neuronal maturation. eLife Sciences. 3:e01201. doi:10.7554/eLife.01201.

Li, Z., D.I. Dordai, J. Lee, and S. Desiderio. 1996. A Conserved Degradation Signal Regulates RAG-2 Accumulation during Cell Division and Links V(D)J Recombination to the Cell Cycle. Immunity. 5:575–589. doi:10.1016/S1074-7613(00)80272-1.

Lindqvist, A., H. Källström, A. Lundgren, E. Barsoum, and C.K. Rosenthal. 2005. Cdc25B cooperates with Cdc25A to induce mitosis but has a unique role in activating cyclin B1–Cdk1 at the centrosome. The Journal of Cell Biology. 171:35–45. doi:10.1083/jcb.200503066.

36 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Ling, J.P., R. Chhabra, J.D. Merran, P.M. Schaughency, S.J. Wheelan, J.L. Corden, and P.C. Wong. 2016. PTBP1 and PTBP2 Repress Nonconserved Cryptic Exons. Cell Reports. 17:104–113. doi:10.1016/j.celrep.2016.08.071.

Llorian, M., S. Schwartz, T.A. Clark, D. Hollander, L.-Y. Tan, R. Spellman, A. Gordon, A.C. Schweitzer, P. de la Grange, G. Ast, and C.W.J. Smith. 2010. Position-dependent alternative splicing activity revealed by global profiling of alternative splicing events regulated by PTB. Nat. Struct. Mol. Biol. 17:1114–1123. doi:10.1038/nsmb.1881.

Love, M.I., W. Huber, and S. Anders. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. 15:550. doi:10.1186/s13059-014-0550- 8.

Lu, R., K.L. Medina, D.W. Lancki, and H. Singh. 2003. IRF-4,8 orchestrate the pre-B-to-B transition in lymphocyte development. Genes Dev. 17:1703–1708. doi:10.1101/gad.1104803.

Ma, S., S. Pathak, M. Mandal, L. Trinh, M.R. Clark, and R. Lu. 2010. Ikaros and Aiolos Inhibit Pre-B-Cell Proliferation by Directly Suppressing c-Myc Expression. Mol. Cell. Biol. 30:4149–4158. doi:10.1128/MCB.00224-10.

Mandal, M., S.E. Powers, K. Ochiai, K. Georgopoulos, B.L. Kee, H. Singh, and M.R. Clark. 2009. Ras orchestrates exit from the cell cycle and light-chain recombination during early B cell development. Nat. Immunol. 10:1110–1117. doi:10.1038/ni.1785.

Monzon-Casanova, E., M. Screen, M.D. Diaz-Muñoz, R.M.R. Coulson, S.E. Bell, G. Lamers, M. Solimena, C.W.J. Smith, and M. Turner. 2018. The RNA-binding protein PTBP1 is necessary for B cell selection in germinal centers. Nat. Immunol. 45:471. doi:10.1038/s41590-017-0035-5.

Nahar, R., P. Ramezani-Rad, M. Mossner, C. Duy, L. Cerchietti, H. Geng, S. Dovat, H. Jumaa, B.H. Ye, A. Melnick, and M. Müschen. 2011. Pre-B cell receptor–mediated activation of BCL6 induces pre-B cell quiescence through transcriptional repression of MYC. Blood. 118:4174–4178. doi:10.1182/blood-2011-01-331181.

Otto, T., and P. Sicinski. 2017. Cell cycle proteins as promising targets in cancer therapy. Nature Reviews Cancer 2017 17:2. 17:93–115. doi:10.1038/nrc.2016.138.

Payne, S.R., S. Zhang, K. Tsuchiya, R. Moser, K.E. Gurley, G. Longton, J. deBoer, and C.J. Kemp. 2008. p27kip1 Deficiency Impairs G2/M Arrest in Response to DNA Damage, Leading to an Increase in Genetic Instability. Mol. Cell. Biol. 28:258–268. doi:10.1128/MCB.01536-07.

Peco, E., T. Escude, E. Agius, V. Sabado, F. Medevielle, B. Ducommun, and F. Pituello. 2012. The CDC25B phosphatase shortens the G2 phase of neural progenitors and promotes efficient neuron production. Development. 139:1095–1104. doi:10.1242/dev.068569.

37 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Sasanuma, H., M. Ozawa, and N. Yoshida. 2019. RNA-binding protein Ptbp1 is essential for BCR-mediated antibody production. Int. Immunol. 31:157–166. doi:10.1093/intimm/dxy077.

Schlissel, M., A. Constantinescu, T. Morrow, M. Baxter, and A. Peng. 1993. Double-strand signal sequence breaks in V(D)J recombination are blunt, 5'-phosphorylated, RAG- dependent, and cell cycle regulated. Genes Dev. 7:2520–2532. doi:10.1101/gad.7.12b.2520.

Schmidt, M., S.F. de Mattos, A. van der Horst, R. Klompmaker, G.J.P.L. Kops, E.W.-F. Lam, B.M.T. Burgering, and R.H. Medema. 2002. Cell Cycle Inhibition by FoxO Forkhead Transcription Factors Involves Downregulation of Cyclin D. Mol. Cell. Biol. 22:7842– 7852. doi:10.1128/MCB.22.22.7842-7852.2002.

Shen, S., J.W. Park, Z.-X. Lu, L. Lin, M.D. Henry, Y.N. Wu, Q. Zhou, and Y. Xing. 2014. rMATS: robust and flexible detection of differential alternative splicing from replicate RNA-Seq data. Proc. Natl. Acad. Sci. U.S.A. 111:E5593–601. doi:10.1073/pnas.1419161111.

Shibayama, M., S. Ohno, T. Osaka, R. Sakamoto, A. Tokunaga, Y. Nakatake, M. Sato, and N. Yoshida. 2009. Polypyrimidine tract-binding protein is essential for early mouse development and embryonic stem cell proliferation. The FEBS Journal. 276:6658– 6668. doi:10.1111/j.1742-4658.2009.07380.x.

Shinkai, Y., G. Rathbun, K.P. Lam, E.M. Oltz, V. Stewart, M. Mendelsohn, J. Charron, M. Datta, F. Young, and A.M. Stall. 1992. RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell. 68:855–867.

Spellman, R., M. Llorian, and C.W.J. Smith. 2007. Crossregulation and functional redundancy between the splicing regulator PTB and its paralogs nPTB and ROD1. Molecular Cell. 27:420–434. doi:10.1016/j.molcel.2007.06.016.

Suckale, J., O. Wendling, J. Masjkur, M. Jäger, C. Münster, K. Anastassiadis, A.F. Stewart, and M. Solimena. 2011. PTBP1 is required for embryonic development before gastrulation. PLoS ONE. 6:e16992. doi:10.1371/journal.pone.0016992.

Susaki, E., K. Nakayama, and K.I. Nakayama. 2007. Cyclin D2 Translocates p27 out of the Nucleus and Promotes Its Degradation at the G0-G1 Transition. Mol. Cell. Biol. 27:4626–4640. doi:10.1128/MCB.00862-06.

Thompson, B.J., R. Bhansali, L. Diebold, D.E. Cook, L. Stolzenburg, A.-S. Casagrande, T. Besson, B. Leblond, L. Désiré, S. Malinge, and J.D. Crispino. 2015. DYRK1A controls the transition from proliferation to quiescence during lymphoid development by destabilizing Cyclin D3. Journal of Experimental Medicine. 212:953–970. doi:10.1084/jem.20150002.

Trapnell, C., A. Roberts, L. Goff, G. Pertea, D. Kim, D.R. Kelley, H. Pimentel, S.L. Salzberg, J.L. Rinn, and L. Pachter. 2012. Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc. 7:562–578. doi:10.1038/nprot.2012.016.

38 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Vuong, J.K., C.-H. Lin, M. Zhang, L. Chen, D.L. Black, and S. Zheng. 2016. PTBP1 and PTBP2 Serve Both Specific and Redundant Functions in Neuronal Pre-mRNA Splicing. Cell Reports. 17:2766–2775. doi:10.1016/j.celrep.2016.11.034.

Yang, W., J. Shen, M. Wu, M. Arsura, M. FitzGerald, Z. Suldan, D.W. Kim, C.S. Hofmann, S. Pianetti, R. Romieu-Mourez, L.P. Freedman, and G.E. Sonenshein. 2001. Repression of transcription of the p27 Kip1 cyclin-dependent kinase inhibitor gene by c-Myc. Oncogene. 20:1688–1702. doi:10.1038/sj.onc.1204245.

Yoon, M.-K., D.M. Mitrea, L. Ou, and R.W. Kriwacki. 2012. Cell cycle regulation by the intrinsically disordered proteins p21 and p27. Biochemical Society Transactions. 40:981–988. doi:10.1042/BST20120092.

Yüce, Ö., A. Piekny, and M. Glotzer. 2005. An ECT2–centralspindlin complex regulates the localization and function of RhoA. The Journal of Cell Biology. 170:571–582. doi:10.1083/jcb.200501097.

Zeman, M.K., and K.A. Cimprich. 2013. Causes and consequences of replication stress. Nature Cell Biology. 16:2–9. doi:10.1038/ncb2897.

Zhu, A., J.G. Ibrahim, and M.I. Love. 2019. Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences. Bioinformatics. 35:2084–2092. doi:10.1093/bioinformatics/bty895.

39 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Figure Legends

Figure 1. Lack of PTBP1 and PTBP2 blocks B cell development.

a, Splenocytes pre-gated on live (eFluor780-) cells. Numbers in plots show proportions

of B cells from mice of the indicated genotypes. b, Numbers of B cells (B220+CD19+)

identified as shown in a per spleen from mice with the indicated genotypes. Data shown

is from one representative out of two independent experiments. c, Gating strategy based

on cell-surface markers for developing B cells from bone marrow cells pre-gated on dump

(Gr-1, CD11b, NK1.1, Siglec-F and F4/80)-negative live (eFluor780-) cells. Full gating

strategy is shown in Supplementary Fig. 1f. d, Numbers of developing B cells in the bone

marrow (2 femurs and 2 tibias per mouse) of mice with the indicated genotypes. b, d,

Bars depict means, each point represents data from an individual mouse and P-values

were calculated by one-way ANOVA with Tukey's multiple comparisons test. Summary

adjusted p value *<0.05, **<0.01, ***<0.001, ****<0.0001.

c, d, Data shown is from one representative out of three independent experiments with 4

to 5 mice per genotype (except for Rag2-/- mice from which one or two mice were

included) carried out with the same mouse genotypes shown except for the

CD79aCre/+Ptbp1+/+Ptbp2+/+ mice which were only included in the experiment shown here

(three CD79aCre/+Ptbp1+/+Ptbp2+/+ mice).

Figure 2. Cell cycle progression in Pro-B cells.

a, EdU and BrdU sequential labelling experimental set up to distinguish early, late and

post S-phase cells. b, Flow cytometry data of the different stages of S-phase in pro-B cells

(B220+CD19+IgD-surfaceIgM-intracellular-Igµ-CD43high) identified as shown in

Supplementary Fig. 3a. Numbers shown are proportions of cells. c, Percentages of pro-

40 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

B cells in different S-phase stages determined as shown in a and b. d, Number of pro-B

cells in bone marrow (2 femurs and 2 tibias per mouse) at different S-phase stages

determined as shown in b. e, Flow cytometry data of pro-B cells identified as shown in

Supplementary Fig. 3a and excluding BrdU+-only (cells in early S-phase). Numbers

shown are proportions of cells in the G2/M gate. f, Proportions and numbers (in 2 femurs

and 2 tibias per mouse) of pro-B cells in G2/M identified as shown in e. In c, d, f, bars

depict means, each point represents data from an individual mouse and P-values were

calculated by one-way ANOVA with Tukey's multiple comparisons test. Summary

adjusted p-value *<0.05, **<0.01, ***<0.001, ****<0.0001. b-f, data shown is from one of

two independent experiments.

Figure 3. PTBP1 and PTBP2 absence causes changes in mRNA abundance and AS.

a, Differences in mRNA abundance in pairwise comparisons from pro-B cell

transcriptomes. Shown are Log2Fold changes calculated with DESeq2 (Supplementary

Table 3). Red dots are values from genes with significant differences in mRNA abundance

(padj value <0.05) with a Log2Fold change > |0.5|. Grey dots are values from genes with

no significant differences (padj >0.05) or with a Log2Fold change < |0.5|. Numbers in

plots are the number of genes with increased or decreased mRNA abundance. b,

Differences in AS in pairwise comparisons from pro-B cells. Shown are inclusion level

differences >|0.1| with an FDR<0.05 for different types of alternatively spliced events:

skipped exons (SE), mutually exclusive exons (MXE), alternative 5’ and 3’ splice sites

(A5SS and A3SS, respectively), and retained introns (RI), (Supplementary table 4)

analysed with rMATS. Each dot shows the inclusion level difference for one event. c,

overlaps of genes that have changes in abundance and AS amongst the indicated pairwise

comparisons. d, Heatmap shows z-scores of normalised read counts from DESeq2 from

41 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

each biological replicate for genes that were found with differential mRNA abundance in

any of the three pair-wise comparisons shown in a. e, Heatmap shows z-scores of

inclusion levels from each biological replicate for those AS events that are alternatively

spliced in any of the three pair-wise comparisons shown in b. d, e, Hierarchical clustering

was done on Euclidean distances from z-scores and is shown with dendrograms.

Figure 4. Lack of global enrichment of proliferation-related terms amongst changes

in the transcriptome due to PTBP1 and PTBP2 absence.

a, Selected gene ontology (GO) terms (process and function) significantly (p-value <0.05)

enriched amongst genes that are differentially expressed at the abundance or the AS level

when comparing the transcriptome of P1P2dKO pro-B cells to P1sKO and control pro-B

cells as shown in Fig. 3d and e. Numbers show how many genes a GO term has.

Supplementary table 5 contains all significantly enriched GO process and function

terms. b, Log2-fold changes in mRNA abundance in the indicated pair-wise comparisons

for all tested genes (all genes), genes with increased abundance in S-phase (S) and G2/M-

phases (G2/M). Grey dots show genes with padj>0.05. Red dots amongst “all genes” show

genes with padj<0.05 and a >|0.5| log2-fold change. Red dots amongst S and G2M groups

show genes with a padj<0.05 regardless of their log2-fold change.

Figure 5. CDK inhibitors and CDK activity in P1P2dKO pro-B cells.

a, mRNA abundance of CDK inhibitors in pro-B cells from control, P1sKO and P1P2dKO

mice. Points show DESeq2 normalised read counts from individual pro-B cell mRNAseq

libraries. Bars depict means. DESeq2 calculated p-adjusted values are shown when <0.05

for the indicated pairwise comparisons. b, Intracellular flow cytometry with anti-p27

antibody or control isotype staining detected with an anti-rabbit AF647 conjugated

42 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

secondary antibody. c, Median fluorescence intensities (MFI) from staining shown in b.

d, Intracellular flow cytometry with the indicated antibodies. Numbers show proportions

of gated events. e, Proportions of cells identified as in d. b, d, Each histogram line shows

data from an individual mouse with the indicated genotype. FrB pro-B cells in G0/G1, S

or G2/M phases of the cell cycle were defined by EdU incorporation and DNA staining as

shown in Supplementary Fig. 3g. c, e, Points show data from individual mice. Bars depict

the mean. P-values were calculated by two-way ANOVA with Tukey's multiple

comparisons test. Summary adjusted p value *<0.05, **<0.01, ***<0.001, ****<0.0001.

b – e, Data is from one experiment with four to five mice per genotype. The differences

observed for pT592-SAMHD1 and pS807/S811-RB between control and P1P2dKO cells

were confirmed in an independent experiment where two control and two P1P2dKO mice

were used.

Figure 6. Cell cycle regulators in the absence of PTBP1 and PTBP2.

a, mRNA abundance of cell cycle regulators in pro-B cells from control, P1sKO and

P1P2dKO mice. Genes whose names are in yellow and bold are predicted to have reduced

mRNA abundance due to changes in AS upon Ptbp1 and Ptbp2 deletion. Individual data

points show DESeq2 normalised read counts from individual pro-B cell mRNAseq

libraries. Bars depict means. DESeq2 calculated adjusted p-values are shown when <0.05

for the indicated pairwise comparisons. b, PTBP1 iCLIP and mRNAseq data visualisation.

For PTBP1 iCLIP, x-link sites are shown for all events. Clusters of PTBP1 binding are

shown when found. mRNAseq data from pro-B cells of one replicate per genotype is

shown using sashimi plot visualisation from IGV. Numbers on the left show the maximum

number of reads in the coverage plots. Arcs depict exon-exon junctions detected in

mRNAseq reads. Numbers in arcs show the number of reads for the depicted exon-exon

43 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

junction. Parts of transcript isoforms predicted to be degraded by NMD are shown in

yellow. Parts of transcript isoforms coding for proteins are shown in blue. Exon numbers

of transcript isoforms coding for proteins are shown with E and a number. c, mRNAseq

visualisation as in b from HeLa control cells or cells where PTBP1 and PTBP2 were

knocked down (Ling et al., 2016). d, Human RBL1 normalised DESeq2 read counts from

the two mRNAseq libraries (one control and 1 double knock down) shown in c. e,

Representation of cell cycle mRNA regulon controlled by PTBP1 in pro-B cells and

consequences of Ptbp1 and Ptbp2 deletion in pro-B cells. Depicted interactions of PTBP1

with individual factors are likely to be direct.

SUPPLEMENTARY FIGURES:

Supplementary Figure 1. PTBPs expression in developing B cells.

a, Representation of B cell development in the bone marrow using the Philadelphia

nomenclature (Hardy and Hayakawa, 2003) including Hardy’s fractions (Fr) based on

cell-surface markers (Hardy et al., 1991). b, Identification strategy of bone marrow cells

using cell-surface and intracellular markers. c, PTBP1, PTBP2 and PTBP3 expression

analysed by flow cytometry. Identification of different B cell developmental stages was

carried out as shown in b. d, Geometric fluorescence mean intensity (gMFI) of staining

for PTBP1, PTBP1, PTBP3 and control isotypes as shown in c. Bars depict means. Each

data point shows data from an individual control mouse (CD79a+/+Ptbp1fl/flPtbp2fl/fl). Data

shown is from one experiment with 5 mice. e, Numbers and proportions of B cells

(eFluor780-B220+CD19+) in mesenteric lymph nodes. f, Complete gating strategy of cells

from bone marrow using cell surface markers to identify developing B cells and Hardy’s

fractions B, C and C’ as shown in Fig. 1c. Dump contains Gr-1, CD11b, NK1.1, Siglec-F and

44 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

F4/80. Data shown is from a Ctrl (CD79a+/+Ptbp1fl/flPtbp2fl/fl) mouse. b, f, numbers

indicate proportions of gated populations.

Supplementary Figure 2. B cell development characterisation in the absence of

PTBP1 and PTBP2

a, Cell surface and intracellular staining strategy to define B cell developmental stages in

bone marrow cells from mice of the indicated genotypes. Numbers shown are

percentages of gated events. b, PTBP1, PTBP2 and PTBP3 detection and negative isotype

control staining in different B cell developmental stages identified as shown in a. Data

shown is of a representative mouse out of five for each genotype indicated. Data shown

is from one experiment. c, c-KIT and CD2 expression at different B cell developmental

stages in mice of the indicated genotypes. Identification of different B cell developmental

stages was done as shown in Supplementary Fig. 1f. a, b, c, data shown is representative

from one mouse out of five mice per genotype, except for Rag2-/- mice for which only two

were included. d, c-KIT median fluorescence intensity (MFI) in pro-B cells (FrB and FrC)

identified as shown in c. c, d, Data shown is from one out of three independent

experiments. e, Numbers of developing B cells in bone marrow from two femurs and two

tibias identified as shown in a with a combination of cell-surface and intracellular

staining. Data shown is from one experiment. d, e Bars show means, each point

represents data from an individual mouse, P-values were calculated by one-way ANOVA

with Tukey's multiple comparisons test. Summary adjusted p value *<0.05, **<0.01,

***<0.001, ****<0.0001.

Supplementary Figure 3. Cell cycle analysis in developing B cells.

45 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

a, Gating strategy identifying pre-pro-, pro- and early-pre-B cells in EdU and BrdU

sequential labelling experiments for bone marrow live cells from the indicated genotypes.

b, Early-, late- and post-S-phase gating amongst pre-pro-, pro- and early-pre-B cells

identified as shown in a. c, Proportions of early, late and post S-phase cells amongst pre-

pro-, pro- and early-pre-B cells identified as shown in a and b. d, Numbers of early-, late-

and post-S-phase pre-pro-, pro- and early-pre-B cells identified as shown in a and b in

two tibias and two femurs per mouse. e, Gating strategy to identify cells in G2/M phases

amongst pre-pro-, pro- and early-pre-B cells identified as shown in a. Events shown are

pre-gated excluding cells in early S-phase (BrdU+-only). f, Proportions and numbers (in

two tibias and two femurs per mouse) of G2/M pre-pro-, pro- and early-pre-B cells

identified as shown in a and e. g, Identification strategy of proliferating FrB pro-B cells

amongst bone marrow cells from mice with the indicated genotypes. Mice were injected

with EdU i.p. for one hour before analysis to track cells in G0/G1, S and G2/M phases of

the cell cycle. h, Proportions of FrB developing B cells in G0/G1, S and G2/M phases

defined as shown in g. i, Phospho-histone 3 serine 10 (pH3 S10) staining amongst FrB

pro-B cells identified as in g. j, Percentages of pH3(S10)-positive cells amongst cells with

duplicated DNA amounts (4n) in pro-B cells (FrB) assessed by flow cytometry as shown

in i. a, b, e, g, i, Numbers shown indicate proportions. c, d, f, h, j Bars show means, each

point shows data from an individual mouse and P-values were calculated by one-way

ANOVA with Tukey's multiple comparisons test. Summary adjusted p value *<0.05,

**<0.01, ***<0.001, ****<0.0001. c, d, f, Data shown is from one out of two independent

experiments. h, j, Data shown is from one out of three independent similar experiments.

Supplementary Figure 4. DNA-damage assessment in PTBP1 and PTBP2 deficient

FrB pro-B cells.

46 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

a, FCS-A measurements of FrB pro-B cells in different phases of the cell cycle identified

as in Supplementary Fig. 3g. b, Intracellular staining with anti-pH2AX or an isotype

control. c, MFIs (median fluorescence intensity) from the staining shown in b and fold

changes in MFIs from intracellular staining with pH2AX and isotype control antibodies

between FrB pro-B cells from P1sKO (CD79aCre/+Ptbp1fl/flPtbp2+/+) and control (Ctrl,

CD79a+/+Ptbp1fl/flPtbp2fl/fl or +/+) mice or from P1P2dKO (CD79aCre/+Ptbp1fl/flPtbp2fl/fl) and

control mice in different cell cycle phases from the data shown in b. d, Intracellular

staining with anti-pH2AX FrB pro-B cells in different cell cycle phases. Bone marrow cells

were either stained on their cell surface to identify FrB pro-B cells and fixed before

intracellular staining, EdU and DNA detection (ex vivo) or were incubated with 20 µM

Etoposide in RPMI at 37oC for 3 hours before cell surface staining to identify FrB pro-B

cells (+Etoposide). f, MFIs from staining shown in d and fold changes in MFIs calculated

as described in c. f, Intracellular staining with anti-p53 antibody or with an isotype

control in FrB pro-B cells. Cells were analysed after isolation from the bone marrow (ex

vivo) or after treatment with etoposide as described in d. g, MFIs from the staining shown

in f and fold change in MFI for the different staining as described in c. In a, b, d, and f FrB

pro-B cells were identified in different phases of the cell cycle by EdU incorporation and

DNA staining as shown in Supplementary Fig. 3g. In a, b, d and f, each histogram line

shows data from an individual mouse. In a, c, e and g Each point shows data from an

individual mouse. Bars show means. P-values were calculated by two-way ANOVA with

Tukey's multiple comparisons test. Summary adjusted p value *<0.05, **<0.01, ***<0.001,

****<0.0001. a, data shown is representative from one out of three independent

experiments. b, c, Data is representative from one out of two independent experiments.

d, e, f, g data shown is from one experiment where the treatment with etoposide was

included. Differences in p53 staining levels between FrB pro-B cells from P1P2dKO and

47 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

control mice were confirmed in an independent experiment where two P1P2dKO mice

and two control mice were analysed.

Supplementary Figure 5. Cell sorting strategy of pro-B cells.

a, Cell sorting strategy of FrB pro-B (B220+CD19+IgM-IgD-CD2-CD43highCD25-

cKIT+CD24+CD249+) cells used to isolate RNA and carry out mRNAseq libraries. Dump

contains IgM, CD2 and streptavidin. Prior sorting, bone marrow cells were depleted from

Gr-1, CD11b, IgD, NK1.1, CD3e and Ter119-positive cells.

Supplementary Figure 6. Transcriptome analysis of pro-B cells.

a, Transcriptome correlation in mitogen-activated primary B cells and FrB pro-B cells.

Dots show mean values of TPMs (transcripts per million) from four or five biological

replicates in mitogen-activated B cells and pro-B cells, respectively. Correlation was

calculated with Spearman's rank correlation rho on genes with >= 1 TPM in pro-B cells

and >0 TPM in mitogen-activated B cells (12545 genes). 681 genes had 0 TPMs in

mitogen-activated B cells and >=1 TPM in pro-B cells. b, Proportions of genes with

differences in mRNA abundance for the indicated pairwise comparisons that are bound

by clusters of PTBP1 (iCLIP data) on their 3’UTR (Supplementary Tables 2 and 3). c,

Ptbp2 mRNAseq visualisation with Sashimi coverage plots done with IGV showing one

replicate per genotype. Arches depict reads mapping to two different exons. Numbers on

arches show the number of reads that map to that particular exon-exon junction.

Numbers in black on the left show the maximum number of reads in the coverage plots.

d, Different types of alternatively spliced events analysed with rMATS. e, Proportions of

AS changes between the indicated comparisons that are bound by at least one PTBP1

cluster in the vicinity of the alternatively spliced event (see methods for definition of

48 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

vicinity, Supplementary Tables 2 and 4). f, mRNA abundance of B cell development

regulators in pro-B cells. DESeq2 normalised read counts of the indicated genes. DESeq2

calculated padj values are shown when <0.05. Bars show means. Each dot shows data

from an individual mRNAseq biological replicate.

Supplementary Figure 7. PTBP1 binding to target transcripts

a, PTBP1 binding (iCLIP data) to 3’UTRs. b, Intracellular c-MYC or control isotype staining

of FrB pro-B cells in different stages of the cell cycle identified as shown in

Supplementary Fig. 3g. Each line shows data from one individual mouse. c, Median

fluorescence intensities (MFI) from the staining shown in b. Each point shows data from

an individual mouse. Bars show means. P-values were calculated by two-way ANOVA

with Tukey's multiple comparisons test. Summary adjusted p value *<0.05, **<0.01,

***<0.001, ****<0.0001. Data shown is from one representative out of two independent

experiments.

SUPPLEMENTARY TABLES:

Supplementary Table 1. PTBP1 binding sites (xlinks).

Supplementary Table 2. PTBP1 binding sites (clusters).

Supplementary Table 3. mRNA abundance analysis. DESeq2 results shown in Fig. 3a.

Separate tabs show genes with significant differential (padj<0.05) mRNA abundance with

a log2 fold change > |0.5| for the different pairwise comparisons carried out and also all

the results obtained with DESeq2. Additional tabs shown genes whose transcripts were

bound by PTBP1 clusters at their 3’UTR.

49 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

Supplementary Table 4. AS analysis. Different tabs show inclusion level differences

(IncLevelDifference) shown in Fig. 3b for the three pairwise comparisons carried out. The

first three tabs show significant (FDR <0.05) alternatively spliced events with an absolute

inclusion level difference >0.1. “allresults” tabs show all the results from rMATS. “PTBP1

bound” tabs show those significantly differentially spliced events that were bound in

their vicinity by PTBP1 clusters.

Supplementary table 5. Gene ontology enrichment analysis. Results from gene ontology

enrichment analysis carried out with the groups of genes identified in Fig. 3d, e.

Supplementary Table 6. DESeq2 results for genes shown to have high mRNA expression

levels in S or G2M phases (Giotti et al., 2018) in the three pair-wise comparisons shown

in Fig. 4b.

Supplementary Table 7. Antibodies and reagents used in flow cytometry experiments.

50 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

a Ctrl P1sKO P1P2dKO b CD79a+/+ CD79aCre/+ CD79aCre/+ Ptbp1fl/flPtbp2fl/fl Ptbp1fl/flPtbp2+/+ Ptbp1fl/flPtbp2fl/fl 8 **** 5 +/+ fl/fl fl/fl 10 * **** Ctrl (CD79a Ptbp1 Ptbp2 ) Cre/+ fl/fl +/+ 4 6 P1sKO (CD79a Ptbp1 Ptbp2 ) 10 )/spleen 7 (CD79aCre/+ Ptbp1fl/flPtbp2fl/fl) 48.1 55.4 0.10 P1P2dKO 3 10 4

CD19-BUV737 0 2 3 4 5 0 10 10 10

B220-BUV395 # B cells (x10 0

c Mat. recirc. 10.4 Immature 22.3 14.6 72.8 11.7 26.9 pro 33.5 early-pre pre-pro (FrC) 81.7 (FrC’) Ctrl 41.6 pro CD79a+/+ (FrB) Ptbp1fl/flPtbp2fl/fl 89.5 Late pre 41.8

16.8 13.7 20.4 79.4 14.2 20.1 61.1

76.9 P1sKO 40.7 CD79aCre/+ 22.7 Ptbp1fl/flPtbp2+/+ 83.0

0.12 39.0 94.1 45.9 0.027 53.5 0.77

2.77 P1P2dKO Cre/+ 5.03 CD79a fl/fl fl/fl 56.5 Ptbp1 Ptbp2 99.8

250K 5 4 5 10 0.12 10 78.7 10 62.4 4 13.7 86.2 1.65 0.11 10 200K 4 4 10 10 3 150K 10 3 17.0 -/- 10 3 Rag2 3 10 10 20.9 100K

0 0 50K 25.4 99.8 0 0 IgD- PerCPCy5.5 CD43-bio + SA-BV510 IgM-BV785 FCS-A 0 3 4 3 4 3 4 3 4 5 CD249(BP1)-PE 3 4 5 0 10 10 0 10 10 0 10 10 0 10 10 10 0 10 10 10 IgM-BV785 B220-BUV395 CD25-BV650 CD19-BUV737 CD24(HSA)-FITC

Pre pro (FrA) Pro (FrB) Pro (FrC) Early pre (FrC’) **** d **** * **** **** **** *** ** * 10 15 **** 35 **** 25 **** * **** 20 **** **** 25 5 5 5 5 10 6 15 5 x1 0 x1 0 x1 0 x1 0 4 10

# cells 5 2 5 0 0 0 0

Late Pre (FrD) Immature (FrE) Mature recirc. (FrF) *** **** **** Cre/+ +/+ +/+ *** **** **** mb1 (CD79a Ptbp1 Ptbp2 ) **** **** **** 20 6 * **** 12 *** +/+ fl/fl fl/fl **** **** Ctrl (CD79a Ptbp1 Ptbp2 ) **** **** **** * **** 15 **** P1sKO (CD79aCre/+Ptbp1fl/flPtbp2+/+) 6 6 6 4 8 10 P1P2dKO (CD79aCre/+ Ptbp1fl/flPtbp2fl/fl) x1 0 x1 0 x1 0

# cells 2 4 -/- 5 Rag2

0 0 0

Figure 1 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Ctrl CD79a+/+ P1sKO CD79aCre/+ P1P2dKO CD79aCre/+ a b Ptbp1fl/flPtbp2fl/fl Ptbp1fl/flPtbp2+/+ Ptbp1fl/flPtbp2fl/fl 5 EdU BrdU 10 -1h 40’ -40’ Early Late 2.3 13.7 2.7 13.5 9.7 36.3 4 S S 10

3 10 Post 2 S 10

BrdU 0 82.2 1.7 82.0 1.8 50.2 3.3 BrdU-AF647 2 3 4 5 EdU 0 10 10 10 10 EdU-PB c **** d 12 10 ) ) 60 4 10 * ) **** +/+ 4 * Ctrl (CD79a *** 3 12 3 15 +/+ fl/fl fl/fl ) **** 8 Ptbp1 Ptbp2 ) 5 8 *** Ctrl (CD79a fl/fl fl/fl 8 **** Cre/+ Ptbp1 Ptbp2 ) 40 ** P1sKO (CD79a 2 6 6 8 10 Cre/+ * Ptbp1fl/flPtbp2+/+) P1sKO (CD79a 4 Ptbp1fl/flPtbp2+/+) P1P2dKO (CD79aCre/+ 4 4 20 1 4 5 Cre/+ % Late S % Post S % Early S fl/fl fl/fl P1P2dKO (CD79a 2 Ptbp1 Ptbp2 ) # Pro B (x1 0 2 Ptbp1fl/flPtbp2fl/fl) # Early S pro B (x1 0 # post S pro B (x1 0 0 0 0 0 0 # Late S pro B (x1 0 0 0

Ctrl CD79a+/+ P1sKO CD79aCre/+ P1P2dKO CD79aCre/+ e fl/fl fl/fl Ptbp1fl/flPtbp2+/+ Ptbp1fl/fl fl/fl f Ptbp1 Ptbp2 Ptbp2 * **** * 20 ) 3 5 **** 4 10 Ctrl (CD79a+/+ 15 fl/fl fl/fl 4 Ptbp1 Ptbp2 ) 10 2 P1sKO (CD79aCre/+ 3 10 fl/fl +/+ 10 Ptbp1 Ptbp2 ) Cre/+ 0.18 0.21 11.4 % G2/M 1 P1P2dKO (CD79a 2 5 10 G2/M G2/M G2/M Ptbp1fl/flPtbp2fl/fl) # G2/M pro B (x1 0 EdU-PB 0 0 0 0 50K 100K 150K 200K250K DNA (7AAD) Figure 2 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. a P1sKO vs Ctrl P1P2dKO vs Ctrl P1P2dKO vs P1sKO

4 107 Ptbp2 4 468 4 484 Ciita H2-Aa 3 3 Ptbp2 3 H2-Eb1 Increased Increased Cd74 Increased in P1sKO Cd74 in P1P2dKO H2-Aa in P1P2dKO 2 Capza1 2 2 vs Ctrl vs Ctrl vs P1sKO 1 1 1 0 0 0

−1 Dpp4 Decreased −1 Cd79a Decreased −1 Decreased −2 Ptbp1 in P1sKO −2 Ptbp1 in P1P2dKO −2 Ect2 in P1P2dKO Ect2 −3 vs Ctrl −3 vs Ctrl −3 Iqgap1 vs P1sKO Mpo Log2 fold change Mpo Elane −4 107 −4 758 Elane −4 718 1 100 10000 1 100 10000 1 100 10000 Normalised read counts (base mean) b P1sKO - Ctrl P1P2dKO - Ctrl P1P2dKO - P1sKO 348 44 32 31 43 1165 198 62 103 71 568 92 30 54 66 1.0 1.0 1.0 More More More included included in included in 0.5 in P1sKO 0.5 P1P2dKO 0.5 P1P2dKO vs Ctrl vs Ctrl vs P1sKO

0.0 0.0 0.0 More More More skipped skipped in skipped in −0.5 in P1sKO −0.5 P1P2dKO −0.5 P1P2dKO vs Ctrl vs Ctrl vs P1sKO −1.0 274 43 13 20 17 −1.0 632 90 35 61 32 −1.0 533 108 27 42 39

Inclusion level difference SE MXE A5SS A3SS RI SE MXE A5SS A3SS RI SE MXE A5SS A3SS RI Alternatively spliced event type c P1sKO vs Ctrl P1P2dKO vs Ctrl P1P2dKO vs P1sKO

Different mRNA abundance 181 1078 1117 33 Different 626 148 85 alternative splicing 910 520

d ≠ mRNA abundance e ≠ alternative splicing

−3 0 3 −20 2 Z−score # genes Z−scores (normalised read (inclusion levels) # events counts) Ctrl P1sKO P1P2dKO (cluster) Ctrl P1sKO P1P2dKO (cluster) 182(i)

566 (a) 801 1021 (ii)

455 from 971 (b) 166 (iii) genes

596 611 (c) (iv)

(d) 354 54 (v) 98 (e)

Figure 3 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. a Increased mRNA abundance in P1P2dKO vs Ctrl & P1sKO # genes ag processing and presentation of peptide ag via MHC class II 15 ribosome biogenesis 90 ribonucleoside biosynthetic process 22 rRNA processing 165 neural crest cell migration 14

0 25 50 75 100 % of genes

Decreased mRNA aundance in P1P2dKO vs Ctrl & P1sKO # genes cyclin-dependent protein serine/threonine kinase inhibitor activity 9 lipid catabolic process 123 myeloid cell activation involved in immune response 21 cytokine secretion 24 neutrophil migration 32 response to catecholamine 14 response to lipopolysaccharide 121 positive regulation of protein secretion 127 positive regulation of DNA-binding transcription factor activity 142 positive regulation of glycolytic process 9 positive regulation of MAPK cascade 223 positive regulation of innate immune response 103 regulation of neuron differentiation 362 positive regulation of interleukin-6 production 40

0 25 50 75 100 % of genes

Alternatively spliced in P1P2dKO vs Crl &P1sKO # genes

cytoskeletal protein binding 495

histone H4-K5 acetylation 13

peptidyl-lysine acetylation 89 regulation of neuron differentiation 361 regulation of endocytosis 166 regulation of cytokine-mediated signaling pathway 67

0 25 50 75 100 % of genes b P1sKO vs Ctrl P1P2dKO vs Ctrl P1P2dKO vs P1sKO 4 4 4 3 3 3 Increased Increased Increased 2 in P1sKO 2 in P1P2dKO 2 in P1P2dKO vs Ctrl vs Ctrl vs P1sKO 1 1 1 0 0 0 −1 −1 −1 Decreased Tyms Decreased Decreased −2 in P1sKO −2 Kif2c in P1P2dKO −2 Cdc25b in P1P2dKO vs Ctrl Cdc25b vs Ctrl vs P1sKO −3 −3 Ect2 −3 Ect2

Log2 Fold Change −4 −4 −4 All genes S G2/M All genes S G2/M All genes S G2/M

Figure 4 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

a Cdkn1a (p21) Cdkn1b (p27) Cdkn1c (p57) Cdkn2c (p18) Cdkn2d (p19) +/+ 0.025 <0.0001 Ctrl (CD79a <0.0001 0.0017 <0.0001 0.017 fl/fl fl/fl 1.0 6 <0.0001 10 <0.0001 Ptbp1 Ptbp2 ) 6 6 0.002 <0.0001 P1sKO (CD79aCre/+ 8 Ptbp1fl/flPtbp2+/+) 2 3 3 3 4 4 2 4 6 P1P2dKO (CD79aCre/+ x10 x10 x10 x10

x1 0 0.5 fl/fl fl/fl 4 Ptbp1 Ptbp2 ) 2 2 2 2

Norm. read counts 0 0 0.0 0 0

Ctrl (CD79a+/+Ptbp1fl/flPtbp2fl/fl or +/+) b G0/G1 S G2/M c P1sKO (CD79aCre/+Ptbp1fl/flPtbp2+/+) P1P2dKO (CD79aCre/+Ptbp1fl/flPtbp2fl/fl)

Ctrl 20 *** ) P1sKO 2 15 *** P1P2dKO 10 3 4 5 3 4 5 3 4 5

0 10 10 10 0 10 10 10 0 10 10 10 MFI (x1 0 5 Rabbit IgG + anti-Rab.-AF647 0

Rab Ab + 2ary-AF647 G0/G1 S G2/M ****

) 6

4 **** Ctrl 4 **** P1sKO **** 2

P1P2dKO p27 MFI (x10 3 4 5 3 4 5 3 4 5 0 10 10 10 0 10 10 10 0 10 10 10 0 p27 + anti-Rab.-AF647 G0/G1 S G2/M

G0/G1 S G2/M Ctrl (CD79a+/+Ptbp1fl/flPtbp2fl/fl or +/+) e Cre/+ fl/fl +/+ d 100 33.1 98.7 79.5 P1sKO (CD79a Ptbp1 Ptbp2 ) 80 Cre/+ fl/fl fl/fl 60 P1P2dKO (CD79a Ptbp1 Ptbp2 )

Modal 40 20 0 Ctrl **** 100 **** 100 95.1 **** ** **** 80 16.4 66.1 60 1 RB +

Modal 40 20 50 0 P1sKO *** 100 99.0 80 86.9 35.7 60

Modal 40 % pS807/7 1 0 20 P1P2dKO G0/G1 S G2/M 0 3 4 5 3 4 5 3 4 5 0 10 10 10 0 10 10 10 0 10 10 10 100 pS807/811 RB-AF647 **** **** **** **** 100 50 80 9.65 96.4 49.4 60

Modal 40 * 20 0 Ctrl 0 100 % pT592 SAMHD1 + G0/G1 S G2/M 80 3.99 95.8 55.5 60

Modal 40 20 0 P1sKO 100 80 55.8 97.6 8.72 60

Modal 40 20 P1P2dKO 0 3 4 5 3 4 5 3 4 5 0 10 10 10 0 10 10 10 0 10 10 10 pT592 SAMHD1 + anti-Rab.-AF647

Figure 5 bioRxiv preprint doi: https://doi.org/10.1101/769141; this version posted September 14, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a Foxo1 Myc Ccnd2available underBtg2 aCC-BY-NCRbl1 4.0 (p107) InternationalCdc25b license . Ect2 Kif2c Kif22 <0.0001 <0.0001 <0.0001 10 <0.0001 16 6 <0.0001 15 0.0002 <0.0001 2 12 <0.0001 3 <0.0001 <0.0001 15 <0.0001 0.001 <0.0001 8 <0.0001 2 <0.0001 <0.0001 <0.0001 12 9 <0.0001

3 3 3 4

3 4 4 3 6 3 4 10 10 2

x10

x10 x10

x10 x10 x1 0 x1 0 x1 0 x10 8 1 6 4 1 2 5 5 1 2 4 3 Norm. read counts 0 0 0 0 0 0 0 0 0 Ctrl (CD79a+/+Ptbp1fl/flPtbp2fl/fl) P1sKO (CD79aCre/+Ptbp1fl/flPtbp2+/+) P1P2dKO (CD79aCre/+Ptbp1fl/flPtbp2fl/fl)

Rbl1 (p107) b 3 PTBP1 clusters Kif2c iCLIP 0 2 PTBP1 2 PTBP1 x-link sites x-link sites 0 0 [0 - 660] 619 593 [0 - 1018] 956 698 Control 6 Control

1027 [0 - 654] 565 527 mRNA [0 - 1115] 705 seq P1sKO 9 P1sKO 13 5

563 300 [0 - 605] 375 [0 - 362] 264 26 19 34 P1P2dKO 6 P1P2dKO 7 8 PC PC NMD NMD E7 E8 A5’SS (NMD-triggering exon) E9 E2 NMD-exon E3 E4

Ect2 5 PTBP1 clusters Kif22 0 4 PTBP1 2 PTBP1 x-link sites 0 x-link sites 0 [0 - 870] 712 [0 - 3057] 1904 14 Control Control [0 - 955] 862 [0 - 2711] 9 P1sKO 1433 P1sKO 19 31 73 9 [0 - 177] 42 [0 - 1396] 40 732 24 P1P2dKO 61 104 P1P2dKO 13 PC PC NMD NMD E2 NMD-exon E19 NMD-exon E20 E3

Human RBL1 (p107) c Human RBL1 (p107) [0 - 294] 209 238 d

11 Control ) 4

3 Control HeLa 3 P1P2 KD HeLa [0 - 119] 105 67 36 PTBP1 and 2 PTBP2 KD 1

PC Counts (x1 0 NMD Normalised Read 0 E7 E8 A5’SS E9 (NMD-triggering exon)

e PTBP1 & PTBP2 deficient pro B cells pro B cells - little proliferation enhanced entry into S & block at G2

PTBP1 PTBP1 PTBP1 PTBP1 c-MYC c-MYC x p107 x pT592 p107 FOXO1 P P SMADH1 CYCLIN D2 RB P FOXO1 CYCLIN D2 p27 CDK 4,6 Ccna2, Ccne1... RB Cdk1, Cdk2 CDK 4,6 Ccna2, Ccne1... E2F p27 Cdk1, Cdk2 E2F G0 G1 G1 CYCLIN A,E G0 KIF2c PTBP1 CDK 2 CYCLIN A,E KIF2c KIF22 M S PTBP1 CDK 2 TYMS KIF22 M S ECT2 ECT2 x TYMS G2 PTBP1 G2 PTBP1 pT592 P CYCLIN A,B CYCLIN A,B SMADH1 pT592 CDK 1 CDC25B P SMADH1 CDK 1 CDC25B x

p27 p27 Figure 6